 The plan for this morning is as follows. We are going to hear reports from each of the 11 breakout groups that took place last evening and yesterday afternoon in an integrated fashion that you'll see on this sheet here. It was up on the overhead for a while going a little bit back and forth between the afternoon major thematic breakout sessions and then integrated within them some of the evening breakout sessions. The ground rules are that each of the speakers will give a five-minute summary of the proceedings of the breakout group and we will then be budgeted with ten minutes. Well, actually it's now since we started late about nine minutes and 45 seconds of discussion, maximum. I've been given one major instruction and that's to make sure that we are done by 11 o'clock so that we can have a nice break for about 15 minutes before we have the wrap-up talk by Maynard and I have every intention of keeping the discussion and the speakers on time. I actually would like to, on behalf of sort of the organizing group and NHGRI, I would particularly like to right now thank the breakout session leaders. I was able to flow through about half of each of the sets of breakout groups. I know that all the breakout group leaders, some of them are working hard into the evening, certainly up this morning, coming up with one-page summaries for each of these that are actively being copied and will be distributed sometime during the morning. We will be handing out all the global summaries from each one. In addition, they've all prepared PowerPoint summaries or overheads along the lines. And so I really do want to thank them because some of them were given five minutes notice of their task and I think all of them act in a heroic fashion and have come through. The only other thing I want to promise you is that after yesterday, all the 11 breakout group leaders have promised that there will be no more quotes from Dan Quayle. Two of them was far more than he deserves. So with that, we will start, this is the order that we will go through and the first breakout group summary will be Mike Snyder on biology. Okay, our two breakout groups covered all of biology and really this covered many, many topics and I was actually hoping the summaries would be here because there's a lot of topics on there that I won't be mentioning since I'll simply highlight some of these. We started with the big picture, the 747 view of biology and listed a number of topics in biology that we thought would be not necessarily solved but which there would be major inroads made certainly in the next 20 years and in some of these areas we could envision very significant advances being made even in the next five. And I've listed these on the first slide here and in all of these areas we think genomics will have a very, very major impact. So some of these topics, that's not it, I just lost a minute there. Okay, we also lost the first one. It was actually building a cell in molecular terms. Building an organism, i.e. the molecular basis of development. Organisms interacting with other organisms both within species and between species. Cells and organisms interacting with their environment and some of these turned into breakout sessions in and of themselves including this one. The molecular basis of behavior, intelligence which is viewed as a controversial area and cognition. Comparative genomics which you'll hear more about in a minute. Variation within a population and certainly the relevance of disease and fitness intersecting many of these different areas. Certainly one very, very important part of many of these different so-called big areas of biology would be dissecting key regulatory networks i.e. the arrows. And we certainly, both groups I believe spent a huge amount of time talking about cis-trans regulation and mapping out transcription factor binding sites and identifying those factors, controlling developments, and we certainly got into quite a bit of detail in each of those. But we also at least mentioned, if not spent a little bit of time on other areas including chromatin, chromosomal domains, imprinting, post-transcriptional regulation, molecule-molecular interactions. And I see we got some redundancy here. But all these different areas in which gene regulation and arrows can be drawn. Along the same lines as part of this we would certainly want to characterize in as comprehensive fashion as possible at certainly least the cellular if not the sub-cellular level, the transcriptome and the proteome for every cell type, ideally for every organism that's being actively studied. One last point that's at the end of this is that one we want to integrate all this information using computational and other methods. And finally I forgot to say but this is all intended to be as quantitative as possible. I think that's really going to be a very, very important area of biology in the future to not only draw these arrows but make them quantitative. I want to have much time to go in this. We did touch on technology in a number of different areas in which technology could help advance many of these areas of biology. And that's all I'll say. We'll see what comes up in the next future sessions. We spent more time on two areas, variation within a population, so intra-species variation, as well as variation between species, so comparative genomics and topics related to that. And certainly, again, I won't go through all of these, but one very important theme that came out of both sections is the fact that we need to sequence genomes much, much faster and much, much cheaper than is currently being done. And again, I think that's the topic of another breakout group. And that was a very common theme, as I say, emphasized by both groups. And then there were a lot of other themes that are stressed. Again, I won't go through all of these because they take too much time, but there was clearly a lot of attention devoted to issues about how does genotype translate to phenotype, both within species and between species. Other issues that we thought actually were some which are fairly amenable to analysis now are things like sexual dimorphism. That's an important topic that could be pursued right now. I think another issue that came out as well that was fairly important on all of this is the issue about correlating whether a genetic variation is really positive for disease, and we need new methods and new ways of trying to understand that as well. I won't say much about comparative genomics, but a lot of the themes were the same. And this will be talked about, I believe, more in the next one. One area that we deem very, very important for helping us understand biology is the so-called parts list. And this can break down into several different units. What are all of the different genes for the species we'd like to study, so model organisms and humans? We'd like to have those genes and have them as accurately annotated as possible, both at the DNA level and also at information levels as well. And there are some examples of databases, as Rick Young pointed out for yeast. They're actually pretty good at at least a first level information transfer, although there needs to be higher levels and ways of getting the data more accessible so that people can really manipulate it. Other areas in which we would like to have things cataloged as accurately as possible, again, for defining the parts list, is all a list of alternative transcripts, lists of protein isoforms, and lists of small molecules, the so-called metabolites, lipids, carbohydrates. With each of these parts, it would be useful to get as much information as possible. And again, I think all these are amenable to genomic techniques as well as proteome techniques. And one thing that was unanimous, I believe, in our discussion session was that proteomics is a logical extension of genomics and should really fall under the purview of the Genome Institute. And so we obviously would like to see funding in this area from the Genome Institute. Okay, but this is a list of the kinds of things, fairly obvious one would like to do with the parts. Last, we just started on this topic and actually we were hoping it would turn into a breakout session, which is infrastructure and how funding agencies such as the Genome Institute could, in fact, enhance research in the biological areas. One thing that was clear throughout is that a lot of these problems are going to need multidisciplinary teams to solve, both at a core level and within research groups. And then we envisioned that it would be very, very useful to have resource development to enhance so collections of reagents as well as collections of information that would be very, very valuable to all the people working on these so-called big topics. In addition to that, we envisioned various cores that would be very, very useful. One would be centralized service cores. And actually many of us, including myself, didn't realize that there are in fact sequencing cores out there that you can use. So maybe there needs to be better dissemination of information so that people can know these services exist and they can tap into those. On top of that, we envisioned interactive cores would be very, very powerful for centralized cores, for example, within a university that people could go in and use very expensive equipment, but do hands-on research with genomics and proteomics biology, and that would have significant impact and obviously stimulate research in the area as well. And one can envision a lot of use for small molecule screening, synthesis and screening as well. Finally, overriding all of this computational aspects, again, and quantitative aspects of all of what I said is going to seem to be very, very important. So that's what we covered in a nutshell. So in some ways, it was essentially a mile wide and a millimeter thick, but I'm happy to take comments. Workshop on biology wasn't emphasized here. There were two technological aspects that we felt in our workshop were extremely important for getting deeper than... Be sure to speak in the mic because there are broadcasts. So usually I'm very loud, I'm sorry. There were two aspects that I didn't see in your summary that made it more than a millimeter deep. And one is that it's extremely important to be able to focus on organisms that have good gene transfer technology because the actual functional testing of a lot of the regulatory elements, which are extremely important for actual multi-cellular biology, can only be assayed with great power in such a system where you could do gene transfer. The other thing was that I think there was a lot of interest in getting more sensitive, single-cell, low-copy number molecular reporters so that one can actually monitor expression well and that this is something extremely important, not just peripheral, but extremely important for being able to decode the function of gene products and the expression patterns in organisms. Thanks. Yeah, that actually, the first item about using model organisms you can manipulate came up in ours and stressed very much by your colleague. Other questions? Just to emphasize, the third line down DNA synthesis at a large scale, this was an idea that I came out of our group that I thought was very exciting for 20 years from now. In other words, if one can now synthesize quite easily 50 base pairs, but if we could make that 500 base pairs, then it would be possible relatively easily to synthesize 10 kilobase things by assembling 500 base pair fragments that had good overlapping sequences. So developing that technique of sequence synthesis, not sequence determination, seemed to me to be a very good idea. It wasn't my idea, somebody else in the group said it. I can think of another way in which that would be used as well for those of you who work on strange microorganisms whose codon bias is suboptimal and you want to express those genes in something that's a little more manipulable, such as expressing candid alpha cans and other things. It's very, very useful to be able to synthesize a gene from scratch. In fact, we actually contract out to do that because it turns out it's much, much cheaper. I just want to make a comment about our state of knowledge. I would estimate that maybe half of the sequence dependent information in the animal genome is devoted to regulatory interactions. I think I would estimate that there might be at the most 50 cis regulatory systems that have been analyzed to some reasonable extent in the animal. That's all we know about, to my estimation. And that is a level of ignorance about something which is so important for the morphological species specific characteristics of all animals that it's almost shocking to think of. And so I would really urge us to focus on making some use of the sequence that we now have to understand the most fundamental aspects of development and evolution, which is how these systems work. Okay. If no other comments, thank you, Mike, and we will move on. One of the evening breakout groups was nominally named comparative genomics that had a subtitle of something along the lines of what does it mean to be human? And it was, from what I heard, an interesting breakout session that Bill will give us a summary about. It's okay if this comes out of my time. You just told me you couldn't filter. Okay. Well, this is my PowerPoint presentation in its entirety. So I'm actually a double or possibly triple five minute wonder. I found out five minutes at 7.25 last night that I was sharing this session and I found out at about 8.05 this morning that I was second on the list and my tradition is to make my overheads while listening to the points as I was hoping for temporal sequence. And then I would like to comment that comparative genomics up there as the summary is revisionist, that the summary that I got statement was what it means to be human, which is what we, and the questions, the specific questions really were trying to address what can genomics tell us about what it means to be biologically human. And I think, you know, as with all experiments, you sometimes learn more from experiments that fail than those that succeed. And so I think one, the main take home message that I would give you from our discussion is this probably isn't a good way to phrase the question in trying to think about that. I think it was general agreement, certainly several people very strong. We felt that there's great value in getting primate sequences, which is was the real focus of the discussion. How can our near relatives inform us about the human condition? But that the term biologically human is so loaded with so many values, many of which have truly have nothing to do with biology, but have to do with our social and metaphysical and concept of ourselves. That it's really, we really should be phrasing it much more toward the biomedical issues and all the ways that we can inform ourselves about issues that are very important to human health rather than focusing on these issues. And part of it is part of the aspects of that is also that we're not in a position at this point or any time at all soon to go from having something on the order of 30 to 50 million differences, which is what we would end up with if we had perfect sequence from both organisms to really relating specific molecular differences at the genome level to specific differences at the phenotypic level, which was what we were trying to address. And thinking about it overnight when I would periodically awake from my nightmares, I would add to that that there's a mispremise that the difference that we really didn't discuss. But the idea that it's only the differences that make us biologically human is false. It's some subset of those genomic differences superimposed on all the things that we share with our close relatives that make us what we are. And it's a composite of that. Plus, of course, environment, everything else. So let me in my remaining two minutes. Yeah, two minutes. Just tell you some things I think we can do addressing the question of sort of some of the interesting things. Sort of getting a little bit of the way toward dealing with interesting things about the differences between us and primates that would be important. One is, and this was one of the first points that was made, the identification of rapidly evolving sequences, particularly rapidly evolving gene products. Gene products may be present in one species and not the other, would be very important because those could be sequences that have been under great positive selection to change and that really are species specific markers and may have been fundamental aspects of the process of speciation within primates. And that it's important not only to look at the nucleotide pair level differences, but also to look at large block differences. And this is a gem, I would say this is a general issue for comparative genomics, looking at chromosomal level, changes of large blocks of DNA duplications, deletions, rearrangements may well be extremely important in understanding the processes that extinguish among species. It was felt that in order to have solid information within primates, since we were told that we could imagine as many species sequence as possible, that it would be important to have at least three, preferably more primate sequence. We didn't talk about which ones were the preferable ones. And that, in fact, for understanding the differences, you really need to know about polymorphisms in both, in all the organisms you're dealing with, so having multiple individuals from a species would be a very important attribute. About the only experiment that we could imagine, or not so much experiment, but data gathering exercise that could amplify on understanding these differences, we felt were the possibility of expression arrays and being able to do comparative expression analysis. But in fact, we were told by some members of the group that, in fact, getting material from non-human primates is, in fact, even more difficult than getting it from human primates. So whether there's any feasibility to this is a question that's really open under current standards. And I think the only other thing, and two other things. One, the grass is always greener on the other side of the hill. So the people who work on flies think that yeast is a phenomenally well-understood organism. It's done. The people who work on mice thinks flies are done. And what I was surprised to learn was the people who work on primates think human is done. OK. And I want to make the point that one of the reasons that there's no any hope anytime soon of really trying to relate these 30 or 50 million differences to phenotype is that functional annotation of the human is just so, or any organism really, is just so primitive. We don't know how to put gene products in pathways, how to describe their functions. Maybe if we do, we can begin to overlay something and imagine how these differences are contributing to the phenotypes that are distinct between us and them. And perhaps one of the last things is that perhaps one of the things that you can at least get out of the exercise, even with our current state of non-knowledge, is that at least for the conserved gene products and the conserved domains, we can at least eliminate them as being candidates for what makes us distinct from a primate. And with that, I think I'll stop and ask for your comments. Thank you, Bill. I can't imagine there's not going to be discussion from what I heard at that breakout. So John Vandenberg from the Southwest Regional Primate Research Center. I strongly concur about the need for sequencing, both genomic sequences and CDNA sequencing, from a variety of tissues, a variety of developmental stages, and a variety of species. But I think it's also extremely important to recognize that those of us who work with non-human primates need a lot of other tools from genomics. We need to develop high-resolution linkage maps. We need radiation hybrid maps. We need back-end sequences. We need to develop physical maps. We need microarray technologies that are specific for non-human primate species. We need a database that can collect all of this information and annotate it and relate it to the information from the human genome, and so on. So sequencing is important, but it's not all that's important. The second point I want to make is to respond to your comment about getting material from non-human primates. I disagree that it's difficult to get material from non-human primates. It certainly is difficult to get material from orangutans, gorillas, and to some extent chimpanzees. There aren't very many of these animals. They're very expensive. They're not used in terminal experiments. We're not even breeding them anymore in captivity. But to focus on old-world monkey species, or for that matter, new-world monkey species, there are abundant animals available. You can get tissues from every developmental stage, any kind of cells you want to get. I should correct myself. I made a mistake there. We're really focusing that part of the discussion on orangs, gorillas, and chimps. But I would argue that if we're really going to understand what it means to be biologically human and the differences between ourselves and non-human primates, we need to focus our attention on those primate species that are amenable to functional genomic analyses and experimental interventions. And that means going to the old-world monkeys rather than to the apes. Is this on? Yeah, just speak to the question here. The funny thing is, if it's difficult. It's not on. It isn't on, is it? It's not. There you go. Clean down. The question is about evolutionary divergence and what causes to be human or anything else. And what groups do you really want to study? It's difficult to get human samples if you're going to do microarrays and among close related primates, or orangs and chimps. I think we agree with that. Now, whether you're appropriate comparison with old-world monkeys, that's debatable. If any of those things are true, can you hear me? Yeah, you just got to hear it in the back. If any of those things are true, in other words, if the appropriate comparison is among close related species, then human primates may be difficult. So maybe there are more appropriate model groups, i.e. close related species, in which you can do these studies, sea urchins, tuna kits, something better. So in other words, is there something better out there? I assume the answer is going to be, depends on the questions you want to ask. The question specifically is variation in gene expression causative to phenotypic changes in development. Keith Crandall from Brigham Young University. It bothered me a little bit when we were talking in this discussion. It bothers me a lot now that we're talking about primates under the auspices of comparative genomics. And if we're really going to do comparative genomics, we have to look at a little broader spectrum of phylogenetic diversity. And this very small clade in evolutionary history isn't going to do it. And if we are interested in looking at what makes a human or what makes any individual species, seems to me that we don't need to sequence 10 different primates, we need to sequence three. We need to sequence humans and its sister taxa, chimpanzee, and an outgroup, a gorilla. And then we need to do that in a number of different organisms across phylogenetic diversity. And that gets you information on what makes a species and it gets you independent replicate events to do that. I should clarify, this was not a breakout. This was just a subset topic under the umbrella of comparative genomics. It was by no means meant to be all encompassing under comparative genomics. Mike, we'll take one more. Well, just to add to that last point, my understanding is there's a whole workshop devoted to the issue of what should be sequenced. So I assume they've discussed this in more detail. Yeah, yeah. OK, thank you, Bill. All right, our next breakout group in the evening under the general umbrella of biology was one on gene environment interactions. Well, I actually wasn't even going to do overhead, but I was afraid that the handouts were going to be here. So I did this at the last minute. I should say that this group was actually conducted by David Cox, and so I am sort of a reporter. I will also say that we were given a series of tasks, some of which we did and some of which we didn't, some of which we didn't do. So let me begin by saying what we did do, and then I will talk about what we didn't do very briefly. The first question that we did talk about was how is it possible to approach the issue of gene environment interaction? And the overwhelming conclusion was that taking huge data sets of both DNA and phenotypic information and just sort of approaching them without hypotheses, ex ante, was not going to work due to lack of power. And so what did emerge was a consensus that there are three different strategies that might be useful. One is to do a strategy where you do gene discovery first and then having found genes that you think may be of interest for the phenotype that you're interested in, then explore it in a more genetic epidemiologic fashion with variations of the environment in a more focused way. This seemed very like what sort of is what's already happening. The second suggestion that was made by Janet Rowley initially was that another strategy to take would be to look at things that are known as environmental factors that affect phenotype and use them as tools to approach the question of genetics. And indeed, we already have some examples of this in the context of infectious disease, knowing that some people are more susceptible than others. And it really seemed that that might be the particularly focusing on infectious disease might be a very effective strategy to get to some of the issues about gene variation and how they interact with the environment. As a parenthetical, Alan Gutbacher made the point that this might also address some justice concerns in the sense that this may allow us to understand particularly some of the infectious diseases that affect developing countries, not so much in terms of developing gene-based therapies, but rather having a better idea of pathogenesis that can then give us better ideas of intervention. It was also suggested that drugs in the nature of pharmacogenomics might also be understood as environmental agents that could be understood in this way. And then the final point that was made is that there really will be a use for the very large mother-of-all databases that I know we're going to hear about pretty soon. But the main point is that although there will be a role for that kind of collection, prospectively done in the best possible way, that nonetheless, that if we rely on that strategy alone to try to explore gene environment interaction, that we will have lost the game. The other major point that was the other major issue that was discussed was the issue of what do we mean by environment? And I think that one of the points that we emerged with is that we need to think much more broadly about environment than is typically thought of in most epidemiologic studies. And here, we thought not only about typical toxins and we've already identified some other environmental stressors like infectious agents and drugs, but also that we need to think about other social factors that are quite important. Giulio Licinio made a point that a profound predictor of depression is loss of a parent before 11 years of age. That is not a part of your typical epidemiologic data set. And so one of the other things that we thought it very important to encourage was greater discussion between social scientists and health services researchers and epidemiologists to sort of define the realm of variables somewhat more broadly. So those were the main issues that we discussed. We were tasked with discussing what can be learned from these studies about human behavior and how can this information be used for good or ill. And I will tell you that that did not come up. And it is so I just want to acknowledge that although that was on the to-do list, that did not get checked off. Thank you very much. Questions? I'd also like to point out that there is one school of thought that says that infectious agents are key to evolution. And so that by studying host-agent interactions, one might get some insight into evolution as well. Actually, that point was made. And I'm glad you made that explicitly in this section, because obviously I think many of us think that's true. I just wanted to add is the other major environmental factor in addition to infectious agents is nutrition and diet. Right. I will say we did talk about that and say that it's actually not the evening session, but in the earlier morning in the medicine section it's really hard to figure out what diet is. But certainly, I mean, without a doubt, diet is a huge factor, as all pediatricians know. It sounds like this is mostly human-centric. And I would think that a lot of these same topics would be relevant to other organisms and microbes and flies and things like that. You know? It never got off the human. But I mean, yes. I mean, what can I say? Yes. She's just reporting the facts. No, I, you know, I. OK, thank you, Ellen. Oh, was someone. Oh, was there someone? I think it was a problem that the two last questions, what do we learn from this and what are its consequences, both beneficial and potentially harmful, didn't get talked about. I don't know whether that was because we didn't have enough time. I don't know whether it's because the people who wanted to talk about it waited too long and there wasn't enough time. I don't know whether in this wonderful interdisciplinary mix it's hard for people who are most excited about doing something to think about why we should be doing it. And the virtue of NHGRI is to think about both things and at least that group didn't seem to be able to do that. So part of the difficulty, there was a slight mistake of positioning that breakout group 100 yards from the pub, which may account why some of the questions got dropped near the end. Was there another one up here? As far as you remember, right? OK, I think we will move on. Thank you, Ellen. We will now shift gears and now go to one of our afternoon breakout sessions on technology. And Rick Myers is going to give a summary of that. OK, good morning. I moderated one group and David Burke moderated the other group on technology. And we thought that our charge was to come up with wish lists. And maybe these could be short-term and long-term and we kind of mixed them all in. So let me just go through. And we actually had. We were modest in our wish list. We got it down to 19 of these things on these. And that's after culling away some of them last night. So actually, our group spent a fair amount of time talking about how to do technology development. What are the problems with the ways of doing it? And how do you fund it? How do you deal with people's careers, et cetera? And part of our wish list was to try to figure that out. We've been talking about this in the genome project really since the beginning, the whole question of alternative career pathways, interdisciplinary, et cetera. And the real issue of developing either techniques or instrumentation. And how do you do that when sometimes the instrumentation is not going to sell large numbers of instruments so it's hard to get companies involved in those, et cetera? So that was part of it. A lot of these are really obvious. Not spend time on some of them, but they're obviously very important. Ways to mutate, knock in or knock out, not necessarily with classical genetics, every gene in every cell, any kind of cell, mammalian cells that we put up there, but it can be any type. And then catalog what happens. Really have that information on that kind of scale. Number three is one that has been worked to death, but obviously very important. How do you measure amounts and quantities and maybe qualitative features of molecules and cells, RNA protein, metabolites, et cetera, that's been mentioned. Number four, ways to sequence. I should never put a number up on this. 50, we're gonna want that to be a hundred or a thousand, okay? I'll never forget. Maybe 1991, early on, 93 maybe, early on in the genome project, a very prominent, very well-known genome biologist whose name won't be mentioned, who's on the East Coast, who does a lot of very large-scale sequencing, who's been in the news a lot, said that we'll sequence the planet, or I think that was the phrase, or the biome, maybe that was the word that was used, and at least I thought he was nuts, but I don't think he's nuts. I think that this concept, the idea that we should get there is the real question, and our group actually had a little controversy because somebody said, why? Why would you wanna do that? And I think the general idea is that if you could, in the next session we'll talk about this, if you could, there's never such a thing as too much data with regard to being able to use this to understand biology, and there are many, many reasons, okay? So ways to do this faster, cheaper, more automated, and then you would use that in not only in sequencing, but also possibly as a genotyping tool itself. I liked in Janet's talk yesterday, the graphs that were going up to 10 to the 15th base pairs, or 10 to the 21st base pairs, with 10 to the 9th different genomes. I mean, I don't know how many species there are, but not that many, probably, at least not on this planet. But you might wanna sequence the same genome, I mean, the genome from different individuals of the same species. Ways to measure epigenetic phenomena. This turned out to be a really big one for this group, and in David's group as well, which is, and how to do this on a genome-wide scale. We actually have an idea how to do this a little bit. We didn't talk much, our group at least, didn't talk much about how you would do all these, although we kept lapsing into that. Our charge was to come up with what would we like to be able to do. But some of these things can be measured, but it was pointed out, it's not just methylation, not even just the histones and the changes, but really on a broad, broad scale, all the way to the point of knowing where the chromosomes are in the nucleus, that kind of architecture could turn out to be very important, but really trying to do this on, you know, for the whole genome. One that I always hoped we could figure out how to do this, but even more so after this meeting, after listening to a number of people, Gary Carpin and just the idea that Evan Eichler, et cetera, to sequence highly repetitive DNA, we always said we'll put that aside, it's not gonna be as important, et cetera. We knew that we had to put it aside because we had to figure out how to sequence the easy, or even the Eucromatin. But what if we had, maybe it's different technologies, what if we really could blast through these millions of base pairs and centromeres and in any other strange sequence, come up with ways of really doing that, and I think people are developing them, but we would like to see that if that could really happen, then we really would have all the parts, and as it was pointed out by a number of people, these parts are important. Harold Riedman on telomeres, for instance, et cetera, and this would also help in looking at other genomes that are weird, too, that are hard to sequence. Having those technologies would probably make a difference there. We don't have to spend all of our money on that, but we probably could do a lot better with some investment in it. This was another big one, alternative splicing. I mean, probably most of you know the numbers here is partly prediction, but a fair amount of experiment is showing that this is unbelievably common, and some genes would have thousands of alternatively spliced, at least theoretically, and some of them probably are being shown likely to have thousands of spliced messages in fact. And so that combinatorial problem is incredibly hard. I don't think anybody really knows how to solve this one, but if we could, it would be great. But what are they, what are the species, and how much are there in each different kind of cell? And of course, the real question, what did they do? This is also, just a reminder, starts and stops. There are alternative promoters and alternative three prime ends as well, so all of those are big questions. Ways to determine which proteins interact with each other on a genome-wide scale that's been talked about a lot. Ways to identify which proteins bind a particular small molecule, so proteomics in that sense, and this is something that seems near term or certainly being developed. Ways to identify and understand the functions of non-coding RNAs, this was talked about yesterday, Barbara talked about it and a few others as well, and this would be something that we could be systematic about and maybe really answer this question, so we'd like to see that. Ways to measure post-translational modifications on a whole cell or genome-wide basis. Some of those were grand too, I guess. We didn't really, David and I were dividing these up last night, but here was, and we had lots and lots of overlap. It was interesting to see, and maybe not surprising, the two groups. So this number one, same thing, sequence it for $1,000, you'll hear about that next, but they pointed out the important point that a bonus would be figuring out all the haplotypes at the same time, not trivial but possible, I guess. To generate a synthetic genome and type it in and get the synthesis of it sort of step-wise, I think that concept. I actually liked Oliver's point about getting bigger pieces, being able to do this routinely. Oligos were $200 a piece when we first started, and we throw them away now because they're so cheap. The idea of being able to do that larger would be great. Generate molecular and cellular movies of intact cells. Actually, our group brought this up as well. The idea of really looking at the 3D and the temporal imaging of live cells and doing again this step-wise and going up in complexity and really being able to do that in a big way. That doesn't sound genomic to some extent, but it really is the synthesis and the integration of being able to use all the genomic information. I don't think we should put that challenge aside for some other institute or other group or other group of people that should be in our purview too. The informational challenge, our group did this as well a little bit, but we didn't quite know what to do with the, and I get their sessions on computation, but the real question is whether you can use artificial intelligence to help assist or help to get to do this to deal with this huge amount of information. We're getting there, Eric, I think. Generate small molecule inhibitors for every protein. So then again, this is similar to the knockouts, but they really emphasize in their group the conditional nature. This is actually a really important point too, and so more tunable than just a knockout. Identify all the regulatory sequences and molecules for that matter in the genome, including epigenetic issues as well. And then the big question, this is actually really an important one. Human genome population, we never are ambitious enough in some ways. Can we really think about this on a very large scale? Lots of populations and lots of individuals in those populations for study for the next millennia, not just for the next five-year plan. Okay, so I'm almost done, but just to say, we've then thought a little bit about how you might put some of these into workshops. This is an overlapping point. We think that Francis ought to, or groups ought to get together, probably this institute, a workshop on this training issue and to really deal with it. We always talk about it and we did, or at least we did throughout the genome project, but we never go away with any answers, I think. And so trying to figure out, this partly has to do with funding mechanisms. Some of those have been developed. It's not like nothing has been done here. It is easier to do this stuff now than it was 10 years ago, but really trying to figure out how to do that for the future. An epigenetics one, obviously. Across disciplinary, this is David's group, workshop on predictive modeling and bringing in physicists and stuff. And we actually talked a little bit about modeling and simulations as well. And then maybe people are tired of it, really to do something. How are we gonna do this 50,000 fold more efficient sequencing? Then finally, last slide. Evolutionary genomics, and actually this was, again, David's group, non-standard model organism genomics. We all know that we wanna do genomics on organisms that we can manipulate experimentally and do, but there are also many other organisms that some kind of information, partial or complete sequence, partial or complete RNA or whatever, could be valuable. And they thought that that would be a good workshop. Obviously, one that's timely and maybe happening or arranged already is the near future one of dealing with vertebrate genomes that are being sequenced. And how is that gonna be the infrastructure for distributing that? Not just the information, but the materials and stuff. And then, David, what, I'm not sure. This communication and technology and public perception of technology, I'm sorry, I don't remember what that. Oh, that's right, okay. And so people who are science writers for instance, maybe, or something like that are people who are teachers or communicators and I think that would be a good one as well. So anyway, that was our two sessions. Thank you, can we get the lights up? Mike, are you moving towards the microphone? I was struck by the absence on the list of wishlist for technology development of things that are directly aimed at studying to attryptional regulation. So there's nothing about developing high throughput ways of studying protein DNA interactions or of, I don't know whether that just didn't come up. No, that was actually, that was, both groups talked about that. Okay, I figured it just, it seems like an important, if you look at protein protein, small molecule and ways of developing, high throughput ways of characterizing this regulatory network that are not, it seems, echoing Eric's point for that we really don't have any good high throughput ways of sort of bringing the genome revolution to assist. Yeah, no, that's a good point and I'm embarrassed that if I didn't put it down there because I'm actually trying to work on some of those things. I don't know whether this is a charge for your group or not, but it strikes me that it's kind of related to a group that I was involved in in the evening involving how do we actually develop therapeutics from all this. And that is that in a world of limited resources, this is a great wish list and it's really wonderful. But in terms of thinking about how you would get all of this funded, it might be a wise idea to think about to what extent you might be, some of these technology, the items of technology which this might be something, things that the private sector might have some interest in and so you could piggyback on funding that the private sector is already putting into some of this. So institutionally think about what role for the public, what role for the public. That's a great point. In fact, the whole discussion of how do you do technology development, at least our group and I think probably David's group, we discussed that a lot and that pairing. In fact, a lot of that has happened and it will only happen I think for some of these, the sequencing being one of them. And the workshop you mentioned, that would be the kind of issue that you would likely drill deep into. Right. But that is a good point. Other questions? Oh wait, Janet, wait, Rick, you're not done. It's almost maybe a generic comment but a lot of what you're suggesting are the kinds of things that people will put in grants for and unfortunately study sections aren't always of a mindset to really take innovative technology and or suggestions and move with them. And I don't know how we really figure out ways to get this kind of innovative forward-looking research funded and that I think is a major problem and NCI and others had a study section for innovative technology but those people were no more attuned to innovation than the standard study section. So you gotta figure that end in because you relate, they do need funding. That's a great point. We talked about this a lot. I've served on study section for genome actually too and one of them, we got technology grants and it was very difficult because you have a bunch of biologists and one or two token, it really was token, technology people who mostly were trying to kill each other and the grant by the way that they were looking at. And so it's a very hard problem and it needs to be done differently. I saw, there are a lot of heads nodding, Janet. I think there's a lot of agreement, this is an issue. Yeah, I guess following up on Janet's comment, looking at this report, I'd like to get rid of the term techno geek. In which report? This is the written one. In one of the reports. I've never seen that term, so I don't know. Let's see. I don't have a pocket protector on. I do have one pen in my pocket and I think. Oh, it's the next report? Yeah, I think there's better terms for such people than techno geeks. I mean. I couldn't agree more, Davidson. Yeah, I don't think Davidson geek at all. Does anyone have a raise of hands? No, I've seen David, nobody else does either. Okay, well speaking of techno geeks. So the next topic, sorry, David. The next topic, which interestingly came up naturally in the technology session, was the notion of getting a thousand dollar genome sequence. And so David Page led an evening breakout group on this topic. Right, so this proposition, let's see, and how do I advance, I just advance. Okay, great. Okay, so yes, this was the proposition that we discussed last night, which was to sequence your genome, sequence my genome, sequence all of our genomes for a thousand dollars a pop, and we did some quick calculations and figured out that this would amount to only a mere six trillion dollars to sequence the entire human population. Which we thought would fit easily within the expanded budget that Francis anticipates. So anyway, and then this, this quickly, this proposition and discussion was quickly relabelled thousand dollar genome. And so we took that as a cue to run the discussion in the format of a sort of nerdy, low budget game show. And but actually quite seriously, as a group, we were very proud. We realized that we were asking the right question. And in fact, I think it became clear in the course of a very animated discussion that this is a great question. It has a way of focusing the mind, among other things, it simultaneously solves and creates all problems associated with human genetic testing. And okay, so, but no, quite seriously, we think it is a great question and I'd urge you all to use it in your courses and follow up workshops and such. So first question was, is it realistic? Well, we have a small techno geek problem up here in the animation, but is it realistic and when? So we first, we needed to put, we like the numbers aspect of this thing. So we started out by pointing out that it's cost order of magnitude, half a billion dollars to establish the reference human genome sequence. And that if one were sequencing additional human genomes today, that you might be able to realistically push towards a cost of sequencing additional human genomes of about one cent per base. We recognized that to get to a thousand dollar genome represents a cost per base of 10 to the minus five cents. And so we needed to quickly calculate the difference between these two and our team of computational biologists figured that this was five orders of magnitude and we decided we coined a new term, which was the Burke. The Burke is an order of magnitude increase in sequencing efficiency. And so we recognize the difference between where we are today and where we need to be for the thousand dollar genome is five Burke's. Now, this was called a Burke because one member of our group whose first name is Dave, not, it's not me, initially suggested that a Burke might last about two years. But when pushed on this, it grew to an Albert limit of five years. So we just want to say that there are two to five years per Burke. Therefore, the five Burke's amount to a period of 10 to 25 years to the goal. Now, I don't know who put this in the report, but the word techno geeks have somehow mysteriously appears here. The techno geeks, and you know who you are, you were well represented in this group. And as a group, the techno geeks expressed considerable optimism in achieving these five Burke's over the next 10 to 25 years, despite the complete failure to specify any technology. Now, there was speculation that the unwillingness to specify technology was in some way related to intellectual property concerns. This was all very serious. And then there was a discussed an interesting short-term goal, which is only about a Burke away. And that's the million dollar genome. And the number of interesting questions that could be addressed, the number of interesting disease-specific questions that might be addressed if you could have a million dollar genome. And that's only by our calculations two to five years away. Okay, how about applications? Well, we're no longer going in animated fashion here. I think the techno geeks have left us behind in research in medicine and beyond. Well, first of all, it's worth thinking about the numbers again. We'd expect something like five million differences from the reference sequence in each of these additional genomes. And this has complete knowledge of those five million differences. I will say we had no meaningful discussion of error rates. It's a little aside. If we knew those five million differences, it would obviously crystallize the challenge of genotype-phenotype correlation in human biology. And in principle, set the stage for a full frontal confrontation of genetic and environmental determinants of phenotype. It would represent an extraordinary tool for association studies and in principle for individualized predictive medicine. Those are the fairly obvious ones. And then a couple of interesting other applications were mentioned. One was genetic matchmaking. You could show your genome report to your prospective mates and sort of match up all the interesting a lot of some recessives and such. Decide whether you want to go ahead. Embryo selection, a topic that I'm sure we don't want to venture too far in, but it was, we were thinking that there could be a kind of bulk discount package offered to potential parents who wanted to undergo IVF, maybe make a dozen or so embryos. So we'd need to multiply it by a factor of 12 for those 12 embryos, plus some additional increment for the additional difficulty of single-cell whole genome sequencing. And then perhaps the question actually comes into sharpest focus in the context of newborn screening. So imagining this as a newborn screening technology sort of brings all the issues into sharpest focus. And of course would immediately arise the challenging task. They would immediately arise and forever after be pursued the challenging task of distinguishing the perhaps 99% functionally insignificant sites among the, what did I say, 5 million. The 99% functionally insignificant sites among the 5 million from the 1% or roughly 50,000 consequential sites. And we would imagine that among the consequential sites there would be a tiny minority with high predictive value. Some of those would be recognized today. For instance, those corresponding to known autosomal recessive disorders. And for other, for the others of the imagined consequential sites there would be a falling curve of predictive values that would slowly shift with time as disease association insights accumulated. And finally it would bring into very sharp focus the challenge in human genotype phenotype correlation, the challenge of combinatoric genetic determinants. We didn't even do any calculations on the potential interactions among these 5 million nucleotide differences. Okay, concerns arising. Well, basically this heightens and brings urgency and focus to all, to every single LC issue that surrounds genetic testing. And I think it actually has, again, I think the question has tremendous value in this context, in focusing the mind. So we've just got the list of concerns that you're all familiar with and I only have a partial list of them here. So the concerns for individuals about discrimination, privacy, and so on. The concerns for groups. Though we, in many cases, take great comfort in the realization of our 99.9% identity one to another and the fact that variation within groups exceeds that between groups with this massive amount of data available, it would, the issue of distinguishing among groups and among, by whatever definition would come into much clearer and more quantitative focus. And then, down at the bottom, we realized that this question also had a way of leading to immediate solutions to problems like world peace, universal healthcare. And so we pointed out that this has a way of focusing the mind around, for instance, the deficiencies now and in the future with respect to resources in genetic education and genetic counseling. We imagined that seven minute encounter between the physician and patient in which the five million nucleotide differences would be explained. And finally then, and most soberingly, the potential for such information to exacerbate inequities in access to healthcare, so I'll stop there. Questions? Oh, certainly there must be questions. Yeah. Just a quick comment that came up in our discussion, which was that sequencing the whole genome, everybody's individual genome fits into Barbara Wald's sort of completeness perspective and this came up a lot in our discussion about how there's a real sort of fundamental difference between having your whole genome sequence and having just even a comprehensive genotype, even if you have a million SNPs and just the way, I mean, we didn't figure this out, of course, but the ways in which that's fundamentally different from our view of sort of comprehensive genotyping of the world, I think it prevails looking for computational challenges for the next 20 years, figuring out how to use that marginal information in a very productive way to make it worth sequencing rather than just genotyping everybody was. Yeah, so I think, right, I think again, part of the one of the outstanding qualities of the question is its completeness and precision and it felt in the course of the discussion as if it almost had the mind-focusing qualities of that old question about whether to sequence one genome. Yes. So I'm surprised you didn't bring up the data strain or for the other geeks, the computational geeks and the techno geeks. There seems to be a massive challenge for just handling the size of data is coming, is there? We're hearing in the next talk. Okay, so somebody else is gonna do it. Yeah, we defer to our computational colleagues, yes. The interesting thing is that if we sequence all the humans and we ask any biologically informative questions, one of the things we don't understand is how we're gonna analyze that polymorphism because there might be better systems or other systems in which to start testing this hypothesis. We don't need to sequence every Tosofla, for example, but for example, there are groups of species in which we had a better genome. We could start to be able to advise both the computational mechanism and other mechanisms to analyze data that we hopefully get on humans. And so I guess one of the problems in all the groups I've attended is that they're human-centric and we shouldn't forget that there are other systems which allow us to have power and tools to analyze human sequences. Yeah, the only comment I'd add was, again, the reason we liked the question was that it helps focus the mind. It provides a kind of boundary condition to the ultimate complete genotypic information that you might obtain at least in one species, species to which I think for better or worse, political forces will always return us. And it's very interesting to think about how much of the information that you might gain from this approach might you gain by other potentially less expensive approaches, and that's the sense in which I think it's a terrific question. All right, we're gonna take one more question and move on. One of the things we talked about, of course, in this meeting was the notion that there'd be a tremendous amount of information of varying levels well-validated. And I just wanted to mention the thought that came up that one of the roles that genome could help with is establishing some sort of standards for validating how predictive measures would be used in the clinical world. Because there'd be a huge amount of information and no clear standards, no comparable FDA for establishing when it's clinically ready. Right, so we got five birds of time to figure that out, yeah. Okay, thank you, thank you, David. So the next, David, David Hauser will give a summary on computation from the two breakout groups. Okay, thanks. So we had a couple of computation breakout groups. I think the first thing we decided is that some of us were techno geeks and damn proud of it. Looking into the challenges that we had and the goals that we set, starting with shorter-term goals, five-year goals, and then moving on to the grand challenges in the five-year goal category. Really, there's been a lot of talk about a reference gene set, and there are many things that have to go into this set. We heard a lot about all the old splices, we want cis-regulatory elements and everything, so a gene is large here. And we also want all the functional information about that. So what can we do from a computational point of view? I think the feeling was that it would be a grand challenge to do gene structure and function prediction and essentially first principles from a mechanistic point of view simulated in computation. That's in the grand challenge category, but in the five-year goal category, we should certainly concentrate on combining evidence, experimental evidence, and evidence from comparative sequence analysis, computationally for human curators to create a reference human gene set and gene sets for the model organisms. We definitely need a good gene ontology for this for the functional aspects of this. And rather than trying to come up, wait for an AI system that will somehow come back and suck this information out of the archival literature, we wanted to really motivate biologists to use gene ontology along with traditional publication of results. So we have to have everybody working together to help annotate this reference gene set. And I think that there was talk about how we could creatively use the World Wide Web to do that. Ideas of collaborative annotation over the World Wide Web are something that I think are a good challenge that we have to flesh out. The obvious problems with that are maintaining some kind of consistency if you take input from so many different experts. So how can the web be used creatively to adjudicate all of these things and create some data set that is universally in a universal language that we can all share, that the data, you can download it as a both data set and analyze it from a bioinformatics point of view, but also from the individual scientists to be able to just looking at a few genes that they get the information they need about those genes. So satisfying a variety of users is a very important thing. And I think there are great opportunities for this kind of innovation that needs to be explored. One of the other points that came up though is never forget, you know, always show the evidence as we want to move towards more synthetic, computed integrations of the data. The biologists always want to see the raw data behind that, the conclusions that you present along with your reference gene data set. And then finally, linking into some of the grand challenges of kind of understanding how it all works together, I think we need to think about genes not only from an ontological point of view, but we need to start thinking about genes from a quantitative functional annotation. And there was a lot of, there was some discussion in the group about what that would mean. So what kind of quantitative features that can we attach to genes so that somebody could think about then using the same resource in the systems, the complex systems, the systems biology challenges that will be coming along. We have to plan ahead for that. So there were many suggestions along that line. Certainly for enzymes you can imagine quantitative things that you could attach, but what about for other genes and is there some universal thing that we could get? Then, as Barbara said yesterday, of course we want to go from just summarizing everything, we want to go from genotype to phenotype and we won't be able to do that unless we also, from an informatics point of view, think about how we're going to capture this phenotype data as well along with the reference gene data set. And this is very, very difficult because it isn't a priori clear of what we need to measure and it's going to be, well inevitably it will be we want to measure everything. So we want a huge, there's a huge number of measurements that come in and then you have this data overload question. So we started to discuss some kind of mathematical reduction or transformation or unification of this data set that might be valuable so that we could have some kind of a shared phenotype data and the expected thing is that everybody's going to be collecting their own different types of phenotype data that will do no interoperability, there no comparability, there's no modularity to the whole global data set of phenotype data and that will inhibit progress in the long run. On the other hand, we didn't have any strong suggestions for some kind of draconian unification of this data that would make everybody happy. I think that's also quite, maybe that's in the grand challenge category. If we go on, the other thing that we spent time on is looking at comparative genomics and we really want a computationally rich comparative genomics. Some concrete problems that we identified were to identify the orthologs and paralogs and reconstruct the tree for each gene family. This is already a challenging task. Models of electric revolution are needed, better models, more realistic models that take into account different kinds of functional selection and context effects. Very, very little about the functional selection in cis-regulatory regions, for example. And we don't even have adequate models for functional selection in coding regions, even though we should think we know a lot about those, the mathematical models are still naive. Then we wanna connect the functional and structural data in with this comparative genomics. So we actually want to be using this methodology to predict cis-regulatory sites and active sites and protein structures, but we wanna connect that data with them to file a genetic analysis, the molecular evolution analysis, and that requires a lot of algorithm development. Of course, we wanna connect it with gene expression, human variation, and so forth. So there's some very creative algorithm developments. It seems clear that if you had, we were talking about how many different, say, mammalian species you would want before you could really use them to annotate the human genome, to find genes and so forth in a comprehensive way. And we talked about 10 and we started getting greedy and thinking about a hundred to really start understanding this region. If you really had a hundred genomes and you were really looking at a specific gene and you had this very deep, deeply branching tree, you could align all of these, you could really see the molecular evolution would be very, very distinct for coding regions and for non-coding regions. So it should, in principle, trivialize the problem of recognizing coding regions, because the signal should be so overwhelming, but there's still problems in that and it's a huge challenge in developing that as a way of functional analysis of the human genome. The grand challenge is, I think, which we really have to address. In our session, we did spend a lot of time on this $1,000 genome question and what strikes me, what actually disturbs me, in addition to the ethical issues that were mentioned, is this vision of having my desktop computer and having something that, a device that goes along with it that can, I mean, the desktop computer already knows my genome and I have devices that go along with it that can take massive amounts of phenotypic information and add that to it. And so I've got this potential for having an enormous amount of data about myself, not only my genetic situation, but my current state of, my current physiological state. And if all of this information is available, somebody's gonna wanna compute on it and make some recommendations. And I see a collision here coming up because the medical establishment is not set up to handle this and we're gonna have a huge, there's gonna be a huge opportunity for computational algorithms to come in and start extracting things from this data in an unregulated fashion. And it's, if the data's available, people are gonna want it and somebody's gonna provide the data. And so if we don't start thinking early on about what kinds of computations we can get out of this data realistically and start doing it, we're in a position of having a wild west of all kinds of, I mean, this thing is plugged into the internet and you already have my doctor and all of this stuff giving you advice. What kind of advice are you gonna be getting from this data? It's a very significant challenge. And then quantitative models of molecular networks and engineered cells for experimental validation of the models was another grand challenge. So really to put in predictive, if we're gonna predict how the cell works or how components of the cell work in a quantitative, fully simulated predictive model, we need experimental validation of that and it seems naive to think that you can do that on a whole cell basis without breaking it down into individual networks and pathways. And to really get that right, we're gonna need to have an experimental platform where we can engineer cells or cell-like objects that can simulate parts of the behavior of life along with to validate these models. And so that interaction, the computation can actually suggest design experiments for that as well. Thanks. Lights on. Questions. Did you guys talk about standards in developing software at all or did you not talk about that? We didn't spend a lot of time on standards per se. That gets a little dry. It was something for a great focus group, like a special workshop or a special situation, but they're needed. Are you aware of other communities of scientists, not particularly biological scientists, I don't think, who are quite concerned about legislation in Congress that would be attached additional intellectual property protection to databases and they have rallied for the past couple of years to defeat that legislation. I think it's a concern now also for biological scientists because biology is data-driven now these days as well. So it may be something just to be aware of in the background because it is a, I know it's a huge concern for other sets of scientists. We did talk about IP issues of databases and algorithms briefly and how we can make sure that the data that we need and the algorithms that we need are available without constraint. All right, we'll take one more and then we'll move on. A rhetorical question and then a real question for the last two sessions. And maybe it's a reality check, so bear with me. Why would we want to do this? Because- Do this mean you're back $1,000, you know? Yeah, I mean, Justice asked a very basic question. It seems that from our experience today, very, very small percentage of the population wants this information. Now, maybe in the future they'll want it if they can do something positive with it. It seems there are a lot of other institutions, however, that would love it. So who's gonna create and generate the demand for it? I can imagine there are health insurers, employers, future spouses, educational institutions, it goes on and on. And the thought of fair use for this as us putting public dollars into doing this for $1,000 rather than giving $1,000 to people for food on a global level. It's just a, I mean, again, I don't really have an answer to it, but I guess at least I would like us to evaluate whether or not that is even a direction we would wanna have in a report. I mean, now I can imagine it for the scientific community, this is going to give us incredible information, perhaps, or maybe we won't be able to manage it. And there's so many other issues that I'm concerned about that have come out. I just question, was that very basic question asked? Yes, absolutely. Bob, you wanted to make a comment? In response to that, I mean, I really see that the driver, this is gonna be the physician, that the physician is not going to wanna treat individuals without this information because it's going to impact how they, what kinds of advice they give, what kinds of drugs they give, and all kinds of other things. And I don't think that it's gonna be the public, or the genome project that is gonna pay for this as much as healthcare, because I think that's where the demand is gonna come. And I think that, I mean, I'm fully sympathetic to the major ethical issues that all this brings on. And we're gonna have to deal with it. I mean, as David says, the prospect of this really focuses the discussion. I just wanna make a brief comment, and I hope we'll get more into this with the next session, but I would put in a heavy dose of skepticism about whether physicians want this kind of information. The example I think is most informative is, we're not really terribly interested in who has a susceptibility to pneumonia. We don't need to be. Fortunately, we've got good drugs for it. And I think, to me, the tremendous gain of the genomic era is gonna be getting us to antibiotics for lots of diseases. We don't wanna know about risks that we can't act on, and we don't need to know about risks that we can deal with when they come very effectively, or that we have universal prevention for. So I think we should be very skeptical about how broadly we want predictive information. Just speaking from the perspective of basic computational biology research, a thousand dollar human genome means a $30 worm or a $3 yeast. And so we were talking about having a multiple alignment of a hundred genomes as if it was a sort of wish. But if this really comes to pass, we can have a thousand yeast lined up. And so the ability to do comparative genomics across a whole file and at once becomes within the budget of even my lab. You can do more than a yeast. I was gonna save this for the medicine session, but it came up in the computational biology, which is very nice. And I would suggest this is a showstopper for all of you. The next time you have any interaction with a healthcare professional, whether it's a doctor, a nurse, your dentist, somebody, a pharmacist, from whom you're picking up a medication that's been prescribed, try this. The government is spending lots of money on the human genome project to improve human health. I have great confidence that you are up to speed on how genomics is influencing the type of care that you are providing to me and my family. And I think if you were to try this with any of your healthcare providers now, you'd have a very interesting session with them, hopefully. And I think this will come true in the medicine section. You know, our healthcare providers don't know how genomics is influencing the healthcare that they're providing today. And I think what we should be aiming for is through the information resources that become available in a variety of ways that these healthcare providers really will have that at hand. And so it isn't, it's really to provide at the point of care information that's relevant for individuals. I'll just share it from Wisconsin. I think that these breakouts are all beginning to slide together, because this is a comment I would ordinarily have made when we got to the LC section, but Karen Rothenberg started it. And nobody else has any questions. One of the, this is clearly a hobby horse. People from the LC breakout know this, because it didn't make it into the notes, but I was on it all over. I've been tremendously frustrated by the work in my own field and its absence of attention to the forces that drive things toward application or not. And the lack of distinction in our discussions between problems that are simply intrinsically kind of cool to talk about and think about and publish about and get tenure about versus problems that really do have implications for society, whether legal or social or whatever. And part of that is determined by whether or not there's a genuine, realistic demand on the part of users or providers to users, whether or not there's a financial force that will drive the suppliers toward either enhancing that demand or just meeting the existing demand, which in turn relates to things like intellectual property issues, regulatory obstacles, et cetera, what the political and regulatory system looks like in terms of whether it slows down or speeds this up and the institutions that exist to kind of discipline the entire field, whether the institutions are present and active or whether or not there's a kind of absence. And I don't just mean absence of federal regulation, I mean at the level of state tort law, at the level of professional society, self-regulation, at the level of social approbation. I mean these are all forces that serve to either constrain or make flourish various kinds of applications. And the absence of real attention to this, both theoretical and empirical, to understand the economics and the politics of these potential applications has led us to make lists of things like employment discrimination, which in fact has never really happened in a big way despite being talked about for how many years? In the same breath as we talk about health insurance, which is actually a very segmented problem because it's very much an artifact of our current system and different people are in very different positions with regard to being screened or individually reviewed. And finally, a complete absence of attention to whether in fact the applications are gonna be mostly generated outside of a health field because for all of our discussion about the medical applications of all this genomic stuff, the biggest applications we've seen so far are in the non-medical fields like forensic DNA analysis in a variety of settings, criminal and civil. And the lack of attention to that and the way in which that interplays with the technology, the funding, the public support, the familiarity, et cetera, I think misleads us in terms of prioritizing our own areas of inquiry. And I'm a big believer in the notion that there's something here that's intrinsically interesting, goes to the nature of humanity, et cetera, but not every topic is equally worthy of an equal degree of concern and attention. And I do think some distinction among them might help us all have more sensible discussions in light of the kinds of materials you're presenting. Thank you. Okay, we're gonna take one more quick comment and then we're gonna move on. Okay, I just wanted to say this discussion is gravitating a lot towards human health and things, which is very important, of course. But actually, a lot of this information, especially a lot of computational information, really goes to basic elements of biology, which a lot of it's certain, of course. But actually, a lot of this information, especially a lot of computational information, really goes to basic elements of biology, which a lot of us in this room really care about. We just care about it from a scholarly fashion. How do we get to be what we are for all organisms and all cell types? And I think these are really basic questions. There's huge numbers of computational challenges associated with these. We're gonna be inundated with all kinds of functional analyses and proteomic analyses and expression analyses that's just piling in already and it's just gonna, it's going to go up many, many Birk units in the next year or two. And I just think there's just a huge number because these are different kinds of data sets. We're gonna wanna integrate. These are huge challenges. And I think David alluded to some of this, but this is just another huge area that seems to be getting ignored as we especially as we move into the next topics. Yeah, don't forget the computational challenges associated with these data sets. That's the bottom line. Okay, we are gonna move on in the interest of time and we will shift gears into issues relevant to medicine and to provide a summary of the medicine breakout groups as Bob Nussbaum. It's hardly a shift we've been at it. Last night, late last night when I was befuddled with sleep, Val Sheffield and I were getting together to talk about the two sessions that we had taken care of and Val suggested that we flip a coin as to who gives the talk, heads I give it and tails he doesn't have to. So that's why I'm here. So the topic was genomic medicine in 2020. How do we get there from here? I think it's sort of interesting paradox that the more highly focused the question is the broader and broader is the discussion of that question. And in some sense, understanding human variation and how it relates to disease was the central focus, but our discussion really covered the gamut of many different areas, which I thought made it a fairly interesting set of sessions. So one key issue was that people were very interested in improving genotyping and genotyping technology making it cheaper. We didn't coin a term like the Burke for cost of getting genotyping down. All I know is that it's too much now and as opposed to people not being willing to talk about different genotyping methods because of intellectual property issues, I just think people didn't talk about it because there's so many different ones out there and they all cost much too much. The issue of whether one should do this by genotyping specific snips versus just broad re-sequencing of individuals where you could capture haptotype information that was brought up and it's something that we really should discuss. I think one area that came up repeatedly was the point that Bonnie just made. Of course, she was in my session so of course she made it there but I think it was also made in Val's session and that is that I think a lot of clinicians particularly are just sick and tired of being told about what the vision of the genome project is and how it's gonna impact them here and there and whatever. They wanna know what's useful now for their patients that they're seeing this morning when their office opens up. There was also a strong concern that there are health claims made for particular genomic information particularly from, for example, the private sector that are very difficult for physicians to interpret and not just physicians, we should really broaden this. This is where we're talking about healthcare providers, various kinds, including nurses, genetic counselors, et cetera and that some effort needs to be placed into improving the not only education but having a resource where people can go to and find out what is the real health relatedness, relevance, significance, validity of a genomics derived discovery. Is it hyper? Is it real? Is it useful? There was a very strong emphasis on the need for much more social science research to figure out how to communicate genetic information to primary care deliverers and how they can then deliver it to their clients, their patients and to measure the effects and the outcomes of those efforts. With a additional targeted studies on the public perception and expectations of what the human genome project is going to provide. And as part of this was what really is the role of disease testing in its application to populations in terms of predictive testing? Is this a condition for which we can intervene? Is this a condition that we can intervene when it becomes phenotypically obvious, et cetera? The other issue that was talked about quite a bit is really in the general area of disease gene discovery. Very strong emphasis on improving phenotyping in human beings. I think the point was made, Les Beeseker made the point that our genome pipeline has gone like this but our phenotyping pipeline has stayed really unchanged for many years. And that we need to improve our ability to do phenotyping in human beings. One challenge that was thrown out was the ability to do in vivo expression profiling perhaps through some sort of molecular imaging. One example that was given was how are you gonna look at changes in gene expression in a beating heart in a living individual that would not involve getting a piece of muscle tissue out making RNA and putting it against the chip. NIH directed repository of samples with clinical data for future research studies. There were two models proposed with pluses and minuses of each. One is to think about a prospective longitudinal all-comber random kind of study, a very large data set where one would genotype and phenotype quite extensively without a specific hypothesis. This was pointed out that for many diseases this would result in significantly reduced power and that therefore also patient groups in which there are disease specific collections. But I think in either case the point was made that this material needs to be made available for researchers to be able to test hypotheses, ask questions about the specific genetic differences between individuals with different diseases or traits that this shouldn't be locked up and made unavailable to some scientist who has a bright idea and that therefore the need for adequate concerns about the privacy and confidentiality of that information and how it can be made available without causing harms to the people that have been willing. And a very strong I think comment was made about how most people really are very interested in participating in genetics research and would like to make that sort of contribution. And also a proposal for using the clinical research centers more effectively as one venue in which to improve phenotyping. This is sort of very much along the lines of everything else that everyone's been talking about that the genome project is providing. Please provide it, just provide it better and faster. We pathways networks for microarrays, interaction maps, pharmacogenomics for orphan drugs and adverse effects, completeness of the transcriptome, rapid generation of animal models to test hypotheses if you find a particular variation how do you know it means anything and to be able to manipulate the genome of a model organism. So, this all feeds in. And then finally, just the last couple of points. The elements of fairness and equity in the use of genomic information for medicine really came out in two main areas. One was the support of development of diagnostics for orphan diseases that are being neglected because they don't represent a major financial impetus. And the other is just to make sure that genomics research, both in the discussion of it and the execution of it, that there's inclusiveness of minority populations in that research. A couple of the other points that we're touched on. The last thing I would just like to say is I thought there was a good idea that was brought out about trying to increase clinical research that use a genomic information. And as to think about something like the extramural K23 awards or perhaps something equivalent in the intramural programs as a way of mentoring physicians in genomics science and clinical research together to generate a better pipeline of people that are trained to use this information for health and medical applications. That's it, thanks. Okay, thank you, Bob. Questions? Your comment about having a source of information that physicians could trust and go to that was outside of the hype in that makes one think a little bit of the medical letter in drugs and therapeutics which is a really trusted source for physicians about medicines. And you wonder if there couldn't be a genetics letter or if somehow you could work with a medical letter to include genetic information in that way. And that was actually specifically brought up by one of the people in our group as such a venue. Just to support the notion of loan forgiveness, I think that genetic counseling programs really need to be supported and that we should be talking about having training grants available for both tuition, remission and stipends for all genetic counseling programs. Small point, we're just listening to the presentation which is very interesting in some of the others. I keep thinking that what we're addressing is not genomic medicine in 2020 but genomic medicine in 2100. In other words, some of this seems to me may take a hundred years rather than 20 years. And I don't know what the full implications of that are other than perhaps we need to sort of coordinate efforts a bit. So we may have the thousand dollar genome available before we have any idea of what it means or how to apply it which may just sort of confuse matters and I'm just wanting to get any thoughts about the implications of that. Well, I think the genome project has been very good over the last 10 years about anticipating problems even before the problems were on our doorstep. So I guess what I'm saying is that we should be working on anticipating these problems whether we're lucky or unlucky enough to have them on our plate. I wanted to say two things. One, very quickly to expand on the issue about how we're going to train the clinical investigators is to recognize that in fact that's going to be a big project because and we had actually talked about bringing people in to sort of retrain them to think genetically because medical schools don't teach you that and colleges don't teach you that and we really need to sort of reframe the thinking. But the other argument that I'm going to make about medicine is an argument actually from biology which is that I think even in this group which I think is incredibly sensitive. The fact is it seems to me that the appropriate stance from the scientific community and from the medical community about genomics ought to be one of extreme humility and it seems to me that the more we've learned the more we learn that we don't know and that in fact that there is not and I remember the time a decade ago when people would hand up with the disk and say you can put this in a computer and then you'll be able to know what the outcome is going to be and I think about the opening frame of Gattaca where they test the blood sample and then they say the guy's going to get whatever it is he's going to get. You know the fact is we can't predict that and we're not going to be able to predict that any time very soon. And I think that we need to understand that the human who is in front of us is a product of a very complex interaction between their own genomes, the genomes of other organisms, environment very complexly-written. As a general pediatrician what we do here at least at present has almost no relevance to what I do in any even remotely direct sense. So I would just say that this group in particular needs to be out front saying that there's tremendous promise here but we need to put it in perspective and the perspective needs to be very broad. And I just keep coming back to, the biology is critical. The biology is critical because only when we understand the vastness of what we don't know can we put this in perspective. And I just wanted to say that in front of the group. I mean, as an LC type, I just want to say the biology is critical. Well, as a biology genetics type I think the LC is critical. Okay, thank you Bob. Seven down, four to go. We're right on the hour, we're doing great. Okay, our next speaker is David Valley summarizing the proceedings of an evening breakout group which I think was titled the mother of all epidemiological studies. Is that fair? So I feel as if to some extent I've been set up because the last three or four sessions really have I think been leading to the topic of our session and I congratulate the organizers for putting it together that way. And I must say that in this string of presenters of David, I think David Pages' topic, if that was the at one end of the most sharply focused question, I think our task was much more towards the open-ended question. But our task for our group was to think about designing a large, even massive longitudinal population-based resource that would include medical health data, phenotypic data, environmental data, and genotypic information. And the idea here would be to try to put this in a realistic framework and to learn how to go from genotype to phenotype incorporating all of the other variables or at least a large fraction of the other variables that we currently are aware of and understand how it plays out in the phenotype of individuals over a long period of time. It's the first part of the discussion of our group and actually this took about an hour I would say, was that it involved the definition of this study and at least my rendition of it is that this was not, this is not aimed at being a gene-finding study but rather a resource to detect and evaluate gene-gene and gene environment interactions that ultimately produce a phenotype that largely were interested in a medical perspective from health and disease. And it would depend on in many ways and complement in many ways smaller, more highly focused case control studies of specific phenotypes, let's say hypertension, diabetes, and the like. And I think with this sort of definition there was considerable enthusiasm but somewhat reserved enthusiasm in the group. Now, obviously we couldn't put this whole study together but here were some of the issues that were discussed as related to design of such a study. It would be a long-term participation study, it would require a long-term participation of subjects, healthcare providers, epidemiologists, and many other people. This is of particularly of great concern because experience with other studies would indicate that the maintaining act of participation is that you get more motivated if you're dealing with a specific disease, in this case the individuals would not be dealing with a specific disease, but there is some experience with other kinds of studies that are more general in which strategies largely interactions, frequent interactions in education could maintain good collaboration by all involved. It would require a large amount of phenotypic data extent and type unspecified and environmental data and we were cognizant of the kinds of quantitative and computational problems that David Hausler mentioned earlier and we leave it to the compu geeks to figure out how to do this. It would require repeated return of the subjects for evaluation so they would not necessarily not be anonymous and some sort of design form, design scheme would be put together in which a board perhaps would serve as a firewall between the subjects and the investigators to protect the subjects privacy, a mechanism to make individuals aware of certain health risks that were important for their immediate health would have to be built in upfront so that when things were identified that were of significant significance to the health of the individuals they could be notified and appropriate measures could be taken. Source of subjects, age of subjects, kinds and numbers of populations, kinds and numbers of environments, avoidance of cryptic stratification as best as possible, genotyping strategies, re-sequencing, haplotype mapping and so forth and many other issues would have to be looked at in very critical and thoughtful ways before such a study could be undertaken and there was a lot of talk about doing power calculations and mathematical modeling to study this. It would require also broad investigator access and there's something else I was supposed to, oh I see, okay. Obviously the study such a project would have a lot of LC issues related to it and if I could summarize it was the sense that these would be difficult issues but possible to manage solutions on the basis of careful forethought and planning. We didn't think that this was a deal breaker. The study would also have a heuristic value that is to say there are similar studies developing in Great Britain where it's already apparent that such a large resource as this would spawn many specific, more highly focused studies that use the resource and that would be both an added benefit and an added expense. So the bottom line was that this sounded interesting but it needs a very, very careful and critical feasibility evaluation by individuals with a broad briefcase of expertise, listed disciplines there and others and it would also be important to obtain input from people who are involved in ongoing much smaller and much more highly focused the long-term population based studies such as Framingham and Haynes and certain studies as I said that are already underway in Great Britain and such feasibility studies would have to consider issues of design, power, phenotyping, data basing and social and investigator access to these data. By that I mean patient access to the data not society in general. And ultimately I guess the decision to go ahead or not would come down to the old what's the bank for the buck, how big of a bank can we expect to get and how much would it cost and would that consideration make it feasible to go ahead? It would be envisioned that this is far beyond the ability of the NHGRI to support and would be a trans-NIH sort of arrangement. And in discussions that came after the workshop the idea was hatched that this might be something that would be worth devoting an entire workshop to in which one assembled all of the relevant expertise and included enthusiasts and skeptics and let them hash it out and come up with some more concrete and long-term thoughtful recommendations. Thank you. I wasn't in the group, otherwise I would have made the same comment. I don't know if there's someone here from Child Health but this is happening already. It's called the National Children's Study of Health, Growth and Development. I'm just looking at my email which Frances probably just got yesterday also. 100,000 kids, it's in the planning stage. It was authorized by Congress in 2000 with no additional appropriated funds. And so probably someone from Gino more to get in touch, maybe you wanna comment? We did talk about it. I was in the session that David just nicely summarized and it certainly is on the list of studies that might be looked at to see exactly how it might feed into this paradigm. The way the Child Health Study was designed primarily is focused on environmental influences but obviously there's an opportunity there to look at genetics and I know there are people in the room who are involved in an advisory capacity on that very study. The limitation from our perspective is it's on children and if you're looking at adult onset conditions you're gonna have to wait about 40 years before the data starts to appear. And so we thought this might be an interesting component of a study but the study would, if it was gonna be fully powered, have to include sampling individuals of all ages. Although the population homogeneity is a major issue, certainly databases in countries like Denmark, Sweden, Finland which have been in existence for long periods of time which have incredible amounts of phenotype data, certain hand medical health data. I believe are now at least either being considered or are beginning to include genotype data on at least certain types of genotype data. And I'm wondering if there would not be a real asset to try and tap into that already existing very powerful data sets in Europe to facilitate and enhance and obviously save a lot of cost in trying to do that here. Now granted the populations are different in composition, I recognize that and therefore there would be variables but it seems to me we don't wanna spend a lot of money rediscovering a wheel that may already be in existence. Yes, we're appreciative, thanks for mentioning that Mike and that topic did come up in the discussion and the point was made that there are countries or societies where the medical care is very much is much more uniform I guess I would say than in this country and much more universal and that they might, there was lots to learn from those studies and how it relate to this kind of study is something that would have to be explored. There were also concerns about heterogeneity of genetic background and environmental experiences. I have sort of a general comment prompted unfairly by your aside that we'll leave this to the computer geeks. There are a couple other things that have been said at the meeting one by Pavel saying give us a list of 20 open problems where you need new algorithms and prompted by Janet's overview of computational biology where she was asked in the question session what new algorithms do you foresee being developed and this was a difficult question for you I wouldn't know how I'd answer it either saying I can't think of any off the top of my head but clearly a lot of the things we're proposing to do are to gather very large data sets that will require new algorithms and what would be nice to avoid is another situation where we need a hero like Jim Kent and David Hausler to stand up and solve a problem at the last minute so probably we should think about better communication mechanisms between the computer geeks and the people gathering these data so that we see the need for the algorithms early. Yeah I couldn't agree more and I think in fairness our group recognized that there would be at least to our ability to understand it there would be no, there is currently no algorithms available that would enable us to deal with the data in the way one would like to do it so we recognize that there would have to be a very major input on the computational side if this study were to fly and if we could use it in the ways that one would like to use it. A brief comment about the child health thing the oversight committee for the child health program is the long range planning committee of the council of which I'm a member and since it's focusing on the environment at the outset it's a great opportunity to include the genetics and I think some dialogue between the institutes would be extremely helpful still very much in the early planning session. Point well taken. I'd just like to mention that in addition to new algorithms you're gonna need some new theory. We are embarking on a very focused study much like this looking at gonorrhea infection in a Baltimore community over a 20 year time span that includes multi-locust sequencing and all the epidemiological data and there's no population genetic theory on how to deal with longitudinal data. I mean there just isn't any to estimate anything. So in addition to developing algorithms we also need to develop some new theory to deal with these new kinds of data. I came here with some obvious things which I thought the genome product should do the next five years, 20 years and the obvious things were known by everybody already and then I came here to learn what the non-obvious would be and actually most of that stays pretty unclear and focused but one thing that did crystallize in the meetings on the medical side is what you're talking about is definitely the phenotype and the repositories. Definitely we know that the numbers we need for hypertension and diabetes and asthma and so on are much bigger than we need in terms of existing cohorts and we have to design these studies. We need a mother of all asthma project genetics and asthma mother of all hypertension genetics project and mother of all epidemiological study. I think that definitely is on the list of studies that might be looked at to see exactly how it might feed into this paradigm. The way the Child Health Study was designed primarily is focused on environmental influences but obviously there's an opportunity there to look at genetics and I know there are people in the room who are involved in an advisory capacity on that very study. The limitation from our perspective is it's on children and if you're looking at adult onset conditions you're gonna have to wait about 40 years before the data starts to appear and so we thought this might be an interesting component of a study but the study would, if it was gonna be fully powered, have to include sampling individuals of all ages. Dave, although the population homogeneity is a major issue, certainly databases in countries like Denmark, Sweden, Finland which have been in existence for long periods of time which have incredible amounts of phenotype data, certain and medical health data. I believe are now at least either being considered or are beginning to include genotype data on at least certain types of genotype data. And I'm wondering if there would not be a real asset to try and tap into that already existing very powerful data sets in Europe to facilitate and enhance and obviously save a lot of cost in trying to do that here. Now granted the populations are different in composition, I recognize that and therefore there would be variables but it seems to me we don't wanna spend a lot of money rediscovering a wheel that may already be in existence. Yes, we're appreciative, thanks for mentioning that Mike and that topic did come up in the discussion and the point was made that there are countries or societies where the medical care is much more uniform I guess I would say than in this country and much more universal and that there was lots to learn from those studies and how it relate to this kind of study is something that would have to be explored. There were also concerns about heterogeneity of genetic background and environmental experiences. I have sort of a general comment prompted unfairly by your aside that we'll leave this to the computer geeks. There are a couple other things have been said at the meeting one by Pavel saying give us a list of 20 open problems where you need new algorithms and prompted by Janet's overview of computational biology where she was asked in the question session what new algorithms do you foresee being developed and this was a difficult question for you I wouldn't know how I'd answer it either saying I can't think of any off the top of my head but clearly a lot of the things we're proposing to do are to gather very large data sets that'll require new algorithms and what would be nice to avoid is another situation where we need a hero like Jim Kant and David Hausler to stand up and solve a problem at the last minute. So probably we should think about better communication mechanisms between the computer geeks and the people gathering these data so that we see the need for the algorithms early. Yeah I couldn't agree more and I think in fairness our group recognized that there would be at least to our ability to understand it there is currently no algorithms available that would enable us to deal with the data in the way one would like to do it so we recognize that there would have to be a very major input on the computational side if this study were to fly and if we could use it in the ways that one would like to use it. A brief comment about the child health thing the oversight committee for the child health program is the long range planning committee of the council of which I'm a member and since it's focusing on the environment at the outset it's a great opportunity to include the genetics and I think some dialogue between the institutes would be extremely helpful. Still very much in the early planning session. Point well taken. I'd just like to mention that in addition to new algorithms you're gonna need some new theory. We are embarking on a very focused study much like this looking at gonorrhea infection in a Baltimore community over a 20 year time span that includes multi-local sequencing and all the epidemiological data and there's no population genetic theory on how to deal with longitudinal data. I mean there just isn't any to estimate anything. So in addition to developing algorithms we also need to develop some new theory to deal with these new kinds of data. I came here with some obvious things which I thought the genome product should do in the next five years, 20 years and the obvious things were known by everybody already and then I came here to learn what the non-obvious would be and actually most of that stays pretty unclear and focused but one thing that did crystallize in the meetings on the medical side is what you were talking about is definitely the phenotype and the repositories. Definitely we know that the numbers we need for hypertension and diabetes and asthma and so on are much bigger than we need in terms of existing cohorts and we have to design these studies. We need a mother of all asthma project genetics and asthma mother of all hypertension genetics project and mother of all epidemiological study. I think that definitely is crystal clear in my mind that these studies have to happen. So we need the techno Greeks, the geeks with technologies, we need the info Greeks for the information, we need the LC geeks for the ethical framework. We at the end in the genome world certainly have a lot to contribute to these studies but obviously we have to remember the genome project I think it's main goal was medical so we have to actually make the medical, the translation of the genome actually help the multiple sclerosis project and so on and for that the cohort and the phenotypes become actually critical that these resources be made available shared by all as we did for the genome. Thank you, we agree. Okay, I think we'll move on. The next evening breakout group was chaired by Kurt Fischbeck on gene-based therapeutics. So I should say I had over two hours of discussion by very lively outspoken group of people and I hope I can do it justice in condensing about six or seven pages of notes down to one and a half pages. They put it on the computer this morning. Yeah. It's like a written I as well. Is it just the same text that's on it? Well, yeah, it's extracted from that. Maybe it must have gone on to a different computer. Back in the morning. So, yeah, I can do it. Leon, and why, what are you guys using, Mark, what? No, okay, so why don't I pull a switch here so we can technically get Kurt what we need. So why don't we'll just switch these around and why don't we pull this computer and go make sure to have the talk on it. And meanwhile, we'll move the LC summary up. So why don't we just pull this and take it in the back room and make sure it has a... I'll be back. So we're gonna start with Wiley, then move on to Leon and talk about LC and then we'll move back to gene-based therapeutics. I'm gonna give a sort of broader overview of some of the topics that came up in our LC sessions and then Leon will talk about some of the sort of overarching sort of foundational questions that need to be addressed. One of the issues that came out loud and clear as a tremendously important issue, something that's essential to accomplishing the goals we all wanna accomplish is the need to really integrate together to create much more solid partnerships than we now have. And it's interesting to me that this theme of multidisciplinary interaction and communication has come up in a lot of other workshops as well. What we wanna add to this discussion is that when LC, basic science and clinical science try to work together as partners, we need to recognize how multidisciplinary we are. In fact, each one of those domains is multidisciplinary. And we come from different cultures, therefore we use words in different ways. Dialogue is going to be uncomfortable. We're gonna disagree sometimes about fundamental issues. We also acknowledge that negative stereotypes do exist, not in our groups, of course, but the notion of the LC police, the notion of the basic scientist who wants to find something out and is hell bent toward a goal without thinking about LC, those negative stereotypes exist and maybe they're a barrier sometimes and we need to simply talk together and be partners. Tremendous benefits from being partners. It's fair to say, of course, that all good research incorporates LC and all good ethical, legal and social discussion incorporates a solid knowledge of the basic and clinical sciences that it's addressing. So with that thought of partnership as an overarching theme, a few specific points, that it is important to incorporate theoretical as well as empiric research in what we do as we go forward. And again, Leon will talk more about that. A few other issues, diversity. Diversity is a very important issue for us to talk about together. The fact of the matter is that we are different. There's a notion that the 99.9% similarity that genomics is gonna give us will somehow resolve the fact, but there is a fact that we look different, we talk differently, we have different world views, often perhaps seriously different notions about what genomic information should be used to do or what even our research goals should be as we go forward. And we just simply have to be not afraid to discuss differences and incorporate discussion of differences into both our theoretical and empiric research. There are also important and rather significant issues that we need to solve together that have to do with this public, private distribution of effort. Some specific examples there are a need for in-depth understanding as part of the ELSI agenda of what will be the influences of market forces, for example, on issues like intellectual property and as has come up many times, access to benefits for research. In this context, as we think about market forces, what are they, how do they apply to genomics, how are they gonna play out as we have more fruits of the Genome Project is an ongoing discussion about what the appropriate role of public funding is. The idea came up earlier and came up in one of our discussions too, that for example, there may be orphan pharmacogenetic tests that need to be funded publicly. And as we understand what the private sector can do and the influences of the private sector on how things play out, we may have a better idea of what public funding should do and also what regulation should do. There are gonna be major ongoing themes of discussion. I've listed some of the more applied ones and Leon will talk about some of the larger issues, privacy and confidentiality and the implications of these issues for potentials for discrimination are going to be with us. We may hopefully solve some issues and then go on to other issues. The issue of non-medical uses of genetic information, are they a good thing, are they a bad thing, what are they, are gonna stay with us. Another area that came up as a very important issue is the whole issue of conductive research, that there is a tremendous powerful opportunity for partnership between LC basic science and clinical science. We're all frustrated by IRB procedures that are more procedural than substantive. Research ethics in the genomic arena is a very fruitful area for potential investigation. We would also say and add to the list of training issues that LC ought to be part of basic science and clinical science training just as understanding of science should be part of preparing people to do LC. And we'll just end with affirming, as I think all others have, the importance of broad looks at education, how to do it and how to disseminate it to a variety of different audiences. Now, I'll let Leon comment about some of the broader philosophical issues. Our group was charged with taking up some of the large questions and as a result the conversation was rather chaotic and it events the usual tensions between those who are really interested in the large philosophical questions and those who think that we really should get practical and address the practical ethical questions. This summary which you have in the handout is but a partial summary and partial both in the sense of incomplete and partial in the sense that it reflects the perspective of at least the summarizer. One of the points I think that was most important I think to make is the need to address the larger social and ethical issues connected with the genome project for a number of reasons. First the public's disquiet goes beyond although it includes the important issues of privacy and discrimination but there's something else about the genome project that troubles people and makes them realize this is really of great significance and that these questions that the public has are not solely due to ignorance of genetics or superstition and second that the LC type questions do go beyond how they affect the progress of the genome project. There are other goals and goods beyond health and knowledge though those are very important human and one might even say moral goods and also that the task of the LC part of this project should go beyond simply questions of compliance and dealing with the side effects. It seems to me one could make a theoretical argument that a crucial part of the knowledge that we seek about the human genome is the significance of genomic knowledge and technology in human lives. Let us say it's the genome of this particular species that has to take cognizance of the meaning of knowing its own genome and if you talk about moving to higher levels of organization from molecules to cells to organs to organisms finally one has to take cognizance of how the human beings individually and collectively assimilate this knowledge and what difference it makes for their lives and that seems to me a theoretical question a question of knowledge that is part of the LC should be part of the LC charge. Some we then ask well what kinds of large questions and here this is just a partial list. A number of people talked about the importance well let me say the large questions would be that might be insufficiently addressed up till now some of them concern questions related simply to the acquisition of knowledge and others with respect to the uses of techniques for intervention. We talked almost exclusively about the former strangely enough. One set of questions the meaning of genomic knowledge for human self understanding for individuals for families and genealogies about the question of race and about the question of the self understanding of the human species. There were questions about what it means to live with knowledge of disease risk factors and several people worried about what it means in fact existentially to have presented to one the risks of that particular disease especially when there is no cure but possibly just behavior modifying changes that could reduce but not eliminate the risk and what if the genome project discloses that we all have what risk factors for 10, 15, 20 different diseases all of which require changes in our behavior and none of which produce a cure. Question is what does it mean existentially for people to live with that kind of knowledge? Some people wanted to talk about well what really is the ultimate goal of this genome project not just in the narrow sense but in large what are we really trying to accomplish here through the use of this genetic knowledge and then briefly and only briefly we touched on the question about the boundary between therapy and enhancement and is there a reason to start discussion about genetic enhancement or is this really far-fetched. Then questions about making genomic knowledge effectively beneficial beyond the information this was touched on in the plenary sessions yesterday but a number of people in this group especially those who are intimately connected with patient care are concerned about the obstacles cultural, economic, questions of access to care, problems of limitations of education, certain kinds of just behavioral unwillingness to take advantage of the knowledge and it seems to me some systematic study in the LC program about these kinds of obstacles and how they might be addressed it seems to me would be very important if this genomic knowledge is in fact to bear fruit for human benefit. And then the last set of questions we had a bit of a discussion about the non-medical uses while he has already touched on that but we did set aside a little time for discussion of the outline is mistaken it was about institutional rather than instructional questions. What kinds of institutions and forms need now to be developed to address the emerging LC and policy questions? And we now have an LC program we have various kinds of commissions there but since we can't anticipate all of the large ethical questions one ought to think something about the design of institutions and forms where these questions can in fact be discussed and presidential or national commissions are one thing the grant making programs that the genome project sponsors another but it seems to me a workshop thinking through the question of the design of institutions and the development of public forms for the consideration of these questions. And then finally some specific suggestions for the LC program itself question of where will we get the people trained to do this kind of work and the need to think about the personnel development within the LC part of this genome project. And then some questions parallel to the ones that were discussed with respect to the science is there a need to alter the conventional grant making strategies here in order to get the kind of proposals that one wants to address these larger questions. Thank you. Okay, let me get the lights on and questions. In terms of the information that Leon presented I'd like to maybe propose some amendments and under one A where you say the public's disquiet goes beyond maybe it would be a little more helpful to say something like the public interest goes beyond and then the public interest could both be negative and positive. I'll take it. So it could be. Please correct. Okay, so just so I can finish it. So there could be public disquiet but there could also be a public recognition of rewards or benefits and they should both be explored. Absolutely well taken. And then the same thing under number three. Three A is what are the obstacles? Maybe three B could be what are the effective methods that need to be promoted or that are recognized that could be promoted to again improve access to care and so on. Absolutely. That third point was really both a question of double two parts that are missing here. One is there is need here both for theoretical discussion and empirical research. And then the question is what to do about those obstacles or what to do in fact positively to advance this. It's an oversight but absolutely correct. In front here and then. A question for Leon Cass. I thought the range of issues you presented really very important, very interesting. I was wondering two things you mentioned sort of the meaning of the fact that we'll know our own genomes and so the existential issues. Two sets of issues that really haven't been discussed here to form the meetings in this room. And I'm wondering if you can elaborate a little bit sort of the range of issues that were introduced in the meeting that you had or if there's any kind of consensus but sort of the boundaries of those kinds of issues. Well, I mean it would be very hard to characterize the discussion as an exploration of any one of these in any depth. I mean, there were 18 very articulate people in the room and it had something of the character of the first kind of a discussion of such a group where people were sort of and we didn't, that's what you get when you were invited in a way to say think about the large questions that we haven't thought about yet and everybody is brainstorming. But it does seem to me that a relatively neglected part of this discussion, although some people have alluded to it and what they've written, is what actually does it mean to have, to have predictive knowledge of a clear disease for which there's a remedy, fine. What does it mean to have statistical knowledge about a certain kind of risk that involves alteration of your behavior? Do you wanna know that you've got a 30% chance of getting Alzheimer's disease? And what would you have to do to alter it? And further questions about, well, probably it's a question of living in the shadow of that kind of knowledge. And in some cases that you'll get reassurance because people live in the shadow of those kinds of fears as it is with the family histories and we can perhaps help. But then the further question is what, if in fact physicians, insurance companies and others insist that we provide this kind of information and will people really then have an opportunity to say, I don't want this, I don't wanna live in that kind of knowledge? When it turns out to be cost-effective, in fact, to have this kind of knowledge for the sake of medical treatment. Those are kinds of questions, I think, that are important and are relatively under discussed. I'm very sympathetic to the issue that Leon is discussing. But from the point of view of a basic scientist, there's a kind of a weird disconnect here. In the sense that the reason that we do basic science is to get better understanding of biological processes. There is a deep philosophical faith there that with better understanding of biological processes, one will have more intelligent and useful in specific ways of intervening in human health. It seems to me that part of this problem of being faced with the knowledge of one's disease predisposition and no way to cure it, is remedied by actually allowing the research, which will eventually provide cures to go forward. And I think one of the dangers that we seem to face based on the discussion I hear in this room is of being in such a hurry to provide what's perceived to be output for the medical community that we do it in a way which is premature in terms of actually helping people. I think you've just articulated a powerful reason for the partnership we're talking about. Just a brief comment, Land, regarding the existential impact of living with risk information. The few sociological studies that I'm aware of that have ever looked at the public's ability to understand risk and probabilistic information is incredibly poor. And we live in a world of probabilities and statistics all the time and we understand things very differently than the general public does. And when you go and ask people what it means when the weatherman says there's a 25% chance for rain tomorrow, most people say it means you can't rain. Some say, well, it means it rains a quarter of the day, it doesn't rain the rest of the day. And then the rest of them sort of stroke their chins and say, oh, I never listen to that damn weatherman, he doesn't know what he's talking about anyway. And it's because they can't explain what a 25% risk statement means. And there have been some studies, few in number, that tend to support that. So it really reinforces the very concern that you raise about taking this high power genetic information, probabilistic pre-dispositional testing into a massive population of people about diseases which many are feared and dread by the great majority with information that is solely, at least in some instances, of a probabilistic risk nature and leaving people with that for 20 years, 30 years, 50 years to have to deal with it. And I think these are extremely important areas that need real attention before we leap into the lurch as has just been said prematurely and create far more harm than good with this wonderful technology. Actually, if I could just add a comment to that, I think there's quite a bit of empiric data that suggests physicians have a lot of trouble with probabilistic information as well. And perhaps most of us do accept in very specific technical applications. All right, we're gonna do two more. Let's do the one in back first and then the one in front. Adrienne, go ahead. I appreciate the collection of large questions and the sympathy with which they're being addressed here by a lot of people. I guess one thing that strikes me to add is having, and I think it's almost there, but a question about really what we call progress. What is human progress and is genetic knowledge in what way is genetic knowledge a part of human progress? How do we think of human progress? I think there's an assumption that we know what human progress is and that genetic knowledge is on the way to it, but I'm not sure that's, that might be an assumption to think about. Well, one could say it could be attached really to the question of what is the goal here? And I guess implicit in that is, is this a good goal and why? There's certain tacit assumptions which I think make perfectly good sense. The pursuit of knowledge about ourselves is a great good and a high calling and the attempt to relieve man's estate is also a high good. But there are other goods and there are often prices for pursuing certain goods in a monomaniacal way and therefore one has to sort of think about this thing in the larger human context and not simply in terms of that one is comfortable thinking about it when one is at the bench. Okay, one more question. Your comment about the impact on people's lives of probabilistic information that's given to them and then inferring the need for lifestyle modifications is something that actually has been done for many years in the cancer and cardiovascular areas and there are in fact predictive equations, the Gale equation for breast cancer, the Framingham equations for risk that are available on the web. And perhaps else you should think about expanding a little bit beyond genetics and taking a page from some of these other fields and really establishing a study where you look at what is the impact on someone's life of that information given to them. Good point. Okay, thank you. We will now move back to Kurt and we have fixed our PowerPoint issue and we will now hear about gene-based therapeutics. Is that the first one? Yep, got it. Okay, we're there. Okay. I think this still fits in context. So our group wasn't charged exactly with talking about therapeutics. We had three questions to cover and a fourth that was optional in it. I think in contrast to another group that presented earlier, we cruised through those three assigned questions, got through the fourth optional one and went into a fifth one that we made up ourselves. The first question was how do we make genetic knowledge medically useful? And that has two aspects to it, both diagnostic, which we've been talking about in terms of increasing availability of genetic testing and the implications of that. And then the other topic, which we did spend most of the time on was therapy and therapeutics development. And I think within therapeutics development, there are two issues that came up. One is that therapy is something that's been under-emphasized by the genome community. We all talk about how it would be nice but there hasn't been much down and dirty discussion of how therapy will come about, how we'll be able to apply the genome information in developing therapy. And part of the reason for that is that therapy based on genetics is a relatively new thing. The other issue is the problem with orphan genetic diseases, the hereditary diseases, oftentimes have relatively small markets and that makes it difficult to entice the pharmaceutical industry and biotech companies into working on them. So we have a sense that what we need is protocols for therapeutics development, kind of like pathways into the wilderness or where there are pathways, we need to widen them, maybe pave them. Analogies that came up yesterday of pipeline for therapeutics development. It may not be PC, but I was thinking this morning that there's a lot of oil up there on the North Slope and we have to figure out a way to get it down to Valdez so that people can use it. We need to figure out where the gaps in the pipeline are to open up the pipeline and increase the flow. There are several different aspects to therapy, different kinds, approaches to therapy related to the genome project. And first is gene therapy. And I think it's fair to say that our group was, as a group, we were not agnostic about gene therapy, now let alone in 2020. We felt that there is a need for further development of gene therapy for specific diseases like hemophilia where targeted rational approach to gene therapy is likely to pay off in benefits for patients. And second, and perhaps on a larger scale, is pharmaceutical development. And here, this is an approach that needs to be applied to diseases, including rare diseases that have known mechanism and specific targets. And third, there's a need to identify susceptibility factors. These are genetic predictors, not only for response to drug therapy, but also for response to surgical intervention or behavioral changes. An important point here is that the biggest payoffs in therapeutics development from genomics may be in medical interventions for common non-genetic diseases that we developed based on investigation of genetic diseases. So the second question is, how do we initiate transfer of genetic information to those who will develop the therapeutics? And there's a sense here that we're already doing that very nicely with the public databases that are already available, that the drug companies download from these databases daily and put that information to good use. And more information in those databases will make it more useful for therapeutics development, particularly in terms of phenotype-genotype correlation. And this is something that's come up earlier. Getting those databases increasingly annotated with information about human mutations, the effects of human mutations in specific genes, and the effects of up or down regulation or knocking out an alteration of genes in model organisms. We felt that there is a need, and I think maybe the Genome Institute could play a role here, in fostering interaction between the government and academia and industry in the broad area of pharmacogenetics. And this, we felt, could be a topic for a workshop, perhaps in the next year, to foster that kind of information exchange. And maybe there's a role for funding collaborative studies. There's a need also to improve the access to targets. And here we get into thorny issues of patents and intellectual property. And there may be a role here for us as the genome community to get involved in political action. We, let's see, the third question is what training needs are there? And there is a strong feeling that there is a need to train people, people who can make the connection between genomics and therapeutics. And first on this list is ourselves, that as I was saying, that we haven't talked too much about therapy, we don't know too much about therapeutics development, that we as the genome community need to be better educated in how you make a drug, how therapy is developed. We need to engage the people who do know about doing that kind of thing in our conversations, and we need to educate ourselves better. Second, the industry representatives in our group pointed out that the FDA needs to be educated. That there's not as much genetic expertise at the FDA as one would like, and that we need to engage the FDA and help them to better understand the implications of genomics in their line of work. Something that's come up several times already this morning is the need to educate physicians and other healthcare providers' role of educating genetic counselors and then using genetic counselors to educate others. The point that came up is from Art Baudet, is that if there's a treatment that really works that will catch on pretty easily. There's a strong motivating factor in knowing that if you don't use this test, you may be sued because there's a specific and effective treatment that's based on that test. So as therapy becomes developed, that will start to drive the process of education. We won't have to worry about getting students to sit through lectures about genomics. They'll want to know as students and as healthcare providers what's important in terms of their day-to-day patient management. And then finally, we also talked about the importance of educating the public, and maybe that'll be in the next talk too. Our optional issue, LC issue actually, has been talked about some already, is the social implications, social issues that come out of application of the genome project to diagnosis and therapeutics. And most of what we talked about here were diagnostic issues that have come up already. Maybe just a point to make here that we talked about was the cost of genetic testing is not just the cost of the technique and the technology. There are, again, inter-intellectual property issues here. There is, there are licensing arrangements by universities and companies. There's a question of who should be allowed to profit from what's oftentimes research that's funded by the government. We also recognize the value in cheaper genotyping technology that will increase the availability of genetic testing, the value and the concerns that go along with that. We could say that even at $1,000 of pop large-scale genetic testing will be, or complete genetic testing will be beyond the reach of many people in the United States, let alone people elsewhere in the world. Since that the, so we'll have to be concerned about issues related to genetic testing and how these diagnostic tests will be used. That's been discussed already. The, just again the issue of problems with the patenting system, the sense that the patenting system that we have in place is poorly calibrated and that perhaps, while there is value, recognize value in patenting therapeutic agents, there is questionable value, maybe inappropriate to have patenting of diagnostic gene discoveries for diagnostic use. And an issue that applies to both diagnosis and therapy is a conflict of interest and profit sharing and how that affects the risk to benefit calculation for new therapies. Well, the question we made up ourselves went beyond this, is what is the role of NHGRI in all of this? And we recognize the need to redefine the purpose of NHGRI or at least the extramural part of it. As we go beyond having a complete, as we approach the completeness in the genome sequence, go beyond that to application of the genome sequence in particular application in therapy and application benefiting patients. We felt that there's a need for NHGRI to promote high risk projects and create a fertile responsive environment that can adopt new ideas. Support new technologies we've talked about, new approaches, what will be for the next 20 years, what positional cloning has been for the last 20 years. And finally, and importantly, new people, new investigators. We felt that big science is all well and good, but the important discoveries are most often made by individual investigators working at the bench. Questions, Lee Reich. Thank you for a very thorough report on the breakout group. I had a thought overnight that I didn't bring up last night, so I thought I would bring it up this morning. And that is about the possibility of postdoctoral fellowships for ELSI research and preparation. And I'm thinking of this going two ways. First of all, if one had a PhD degree in philosophy or anthropology or a JD degree, if a lab or a clinic would take us a year or two of intense involvement right in the lab or the clinic, and that way we would get a better feel for what's going on. I don't know how feasible it would be to go the other direction and take a year or two studying philosophy or law after a PhD in molecular biology, but I think those of us who teach in the humanities or law would welcome that kind of interaction as well. Funding mechanisms exist for that already. F32, F33s, and it's possible to go either direction. Another idea, training idea that came up that had a fair amount of discussion was a Marquis-type program for supporting new investigators for picking elite new investigators and supporting them for an extended period of time in seven years or so. One comment that I think relates to the previous presentation and brings it together with the current one is the issue of attributable risk. In other words, genomics at least as it stands today is not gonna give you like a certainty that someone is gonna respond to treatment or have an adverse reaction, but I think that the public is jumping to that too quickly. And so let's say there are drugs and you don't have to even use genomics. There are drugs that cause let's say cardiac side effects and you still give it to a person with heart disease because that's the best drug for that particular person. So you can have a person who has a genetic risk of having a serious side effect to a drug and the doctor may still choose to give that drug to that person knowing the risk and monitoring that. But, and the reverse is true. There are people who may be more likely to respond to a drug, but that doesn't exclude people who don't have that right genotype from being treated with the drug and maybe nothing else works for them. They should be given a chance to try that. And I don't think people understand that very well and I think it could become even a legal issue in the future when certain drugs are approved by the FDA for people with specific genotypes. And what about those who don't have that genotype and are not gonna be optimal responders? Should they be denied the drug altogether? Yeah, so as I was saying, I thought that there was a need for interaction between industry, academics, the government, the FDA in trying to resolve issues of pharmacogenomics like this. One more up here, Paul. I just want to return to a topic that I've read several times during the meeting and that is the topic of large, supporting large scale research versus the support of small investigators. And Kurt mentioned an idea that I think came from Jeff regarding the Markle type fellowship. And I want to put it in a different frame that's perhaps easier to understand to all of us. The Merit Awards have been given off over the last 15 years for investigators who've worked a long time in a field and have become established to do work that might be more risky than they would otherwise do. I think we need to have the same mechanism for young investigators. And we ought to have what we might call a junior Merit Award. Something quite concrete that is judged by how well a person did a PhD and what their first year may be if postdoc is. Maybe they don't even have to have finished their postdoc. But to identify young, bright people and give them seven years of modest research to do what they want. Because I think that's where new ideas come from and most of them anyway. And our study sections can't handle that situation. Thank you, Oliver. Thank you, Kurt. And the last breakout group we're gonna hear from is Genomes for the Public. Karen Rothenberg will give a summary. It's always an honor and privilege to be last when we've been sitting here for almost three hours without a break. But I'll do my best to do this quickly. I'm not necessarily a very rule-oriented person but our group followed our rules. And so I'm just gonna show you the actual questions we were asked. And I actually did a narrative rather than bullets so that means you might have to take a little bit of notes while I speak. Okay, what are the audiences that most need to know genomics? Well, in the ideal world, the answer would be everybody. Everybody would, we would want to know genomics. But I think our group had an interesting priority. And we felt that the most important group to know about genomics are children and their teachers. Starting in elementary school, then moving on through the curriculum in a very determined way in both high school and throughout college. And with teachers really being taught how to teach it as well, that 20 years from now, we wouldn't really need to be trying to do what we're doing now unless, of course, our definition of genomics significantly changes. Of course, healthcare professionals, patients, the media, which we thought was very important, policymakers, institutional review boards, and the judiciary who are going to be asked with increasing frequency to use genetic information for non-medical uses. What does each audience need to know about genomics? Well, obviously, the level of depth and focus depends on what your audience is, but there are some basics that we would want all audiences to know. First of all, the very basic of what is genomics and what is the genome. Very significant is the interaction of the genome and environment in better understanding what is, in fact, that interaction and what, in fact, it makes to make up the human being as a very important principle that came over and over again so that it isn't just something that is relevant to people that see themselves as being at risk for disease but relevant to all of us with respect to our health. The mechanisms of DNA, their applications and the LC issues were all issues that we thought each audience would need to be focused on. Something that's come up again over and over again, and this shows how clever Eric was to put this last because in some way it synthesizes a lot of these themes, is the concept of risk and an understanding of that really not only for the public but for healthcare providers as well, how you explain risk, how you interpret risk, how you assess risk is very important as a concept. And then finally, and again, this is a theme that came up throughout the morning, are realistic expectations. How do you take this information about genomics and put it in a larger context and in perspective as it has an impact not only on healthcare decisions but on decisions in which society may integrate this sort of genetic information into decisions that they make on a daily basis about all other things. Decisions about employment, decisions about custody, decisions about causation, I mean depending on the context that you're in. What are the best ways to reach audiences and what methods work best to increase information retention? Now we were very lucky in our group to have some people that have expertise and experience in trying to provide genomics education and I have to admit quite honestly that there was an incredible level of pessimism about how to get genomic information out to the public. That unless somebody is motivated to wanna hear the information, it is very difficult to be able to retain it. So you have to be very deliberate with respect to the context of which you provide the information so that there has been very negative experiences with respect to providing genetics in with two general healthcare professionals but much better success when you integrate for example cancer genetics in the context of an oncology meeting. Patients that are at risk for diseases generally will tend to be better listeners. In fact, they may know more through the internet than their healthcare provider may know when they come to them and ask them questions but they may know a little bit about a lot and so if they don't have a sort of the depth of understanding that can be sometimes difficult as well to communicate. So the challenge would be then how do we get the general population interested and understanding with respect to genetics, genomics. And I think there was unanimous enthusiasm for this idea that Brad Marcus had. So get your pencils down now because you want to patent this before he patents this. And this is an interactive video game that we will develop for kids. Now, his concern is this might be of more interest to boys than it is to girls. I don't know about the gender implications so I was dreaming about this overnight. I'm gonna come up with a perhaps another addition to this. So stage one of the shootout interactive video game would be you're flying over the DNA double helix to find the exons. By stage two, you're hunting for the mutations that are going to transcribe and translate. By stage three, the body will start smoking and blowing up because you're going to have to deal with it copying and dividing this, it's just DNA demise. And by stage four, this guy in some way is having sex. Now, so totally gender neutral. I don't know whether this would make it all around but I thought we could have a little spin on this and maybe we could even have this done for free rather than as he estimated this costing about $2 million to develop, which is really a pretty much of a drop in the bucket anyway these days. What about Harry Potter? What about thinking about developing something magical? That might be a very interesting idea and we might be able to get somebody to wanna collaborate in some way with that. So we had a lot of enthusiasm but seriously, I think the message here is that just providing information without very serious thought about how you do it is sometimes worse than not providing any information at all. Particularly when you're trying to integrate this to children and in school with science. And this isn't unique to genetics, it isn't unique to genomics. I mean, this is a challenge for all educators so I think that's a very high priority. What do we seek to achieve by informing the public? And again, this is just the very basic question about why we're all here today. I mean, what's the goal ultimately in genomics and why is it so important to us that we have to inform the public about it? So there's a sincere reason, better understanding and appreciating its significance, understanding in the future, in a context of individual health, public health, new drug discoveries, future genetics research and the societal implications. There also is a cynical view that the human genome project to a very large extent was supported by the public and by the taxpayers. And we want to have a continuing understanding and support as we move forward into the next generation. And then finally, whose responsibility is it to inform these audiences? All of our responsibilities. Certainly the Institute, other government organizations, other professional organizations, educational organizations, we put a lot of focus on really bringing higher education and K through 12 educational institutions and really thinking deliberately about this as well. Clearly consumer organizations which have interesting perspectives and a very important role here. And it's not just the responsibility to get the information out, but just as importantly the responsibility to explore innovative ways to continue and improve our understanding. Okay, thank you. Thank you, Karen. Questions, comments? Tim Leshyn from the NHGRI. In your talk, you talked about high school kids, but in your right, if you talk about elementary kids and I assume you mean all of them, I think it's very important to start early and maintain that momentum, but easier said than done, I know. And I think one area you're gonna have to look at are textbooks and how they are written and are they up to date? I'm sure they're not at this point with regard to the Gino. I think we need to think about it. And one last thought I had is with regard to the video game, there's this game SimCity, where you create a city. Why not create a Gino? Yeah. It was not as much fun as having sex, I suppose. I just wanted to clarify the write-up. It says to begin to be educated as early as elementary school, so it's in there. Yeah. You could include sex in SimCity if you wanted to. That must be a California thing. Go ahead. It seems to me that one of the goals, obviously, is to inform the public so that when new technologies come to the forefront, they're willing to accept it. So there's a little bit of brainwashing involved here. And I think the goal of educating, starting at very young ages, is a noble one and should ultimately help, but that's gonna take many, many years. In terms of adults, we're fighting against things like religious beliefs and very deeply ingrained suspicion about genetic modification. And the two most recent examples, we did an abysmal job of selling genetically modified organisms and of stem cells and cloning to the public, and that's clear in terms of what the reactions have been. So what I wanna hear about is how the next time around, when we have issues like this that come to the forefront, are we going to be more effective at convincing the public for their own good that these are at least worthwhile pursuits? Whoa, that's an interesting question you would raise to a law professor. So let me just respond quickly. I think your characterization is problematic. And in fact- Well, it's being extreme, but- That's okay. No, you made the point, though, in a very good way, in a very clear way, because it sounds in our society for good or for bad, it's a pluralistic society, not a paternalistic society. So I think how you frame the educational or information message cannot be laden with, this is what's good for you, but rather to have a sense of open-mindedness as to what the implications may be with respect to your options. And religion for whatever it's worth in this society is a legitimate value that we have, and I don't think, and there are many people in the room who have very strong religious convictions and at the same time feel very strongly about genomics, so I wouldn't pit those against each other. I think we just have to do a better job, which many of us have done, I think, in educating the media and policy makers that a better understanding of the science may help to better gray or blur what people have seen to be the dichotomies. So I think it's all of our responsibility to clarify that much better than we've been able to. Do you think that the issue of genetic modified foods should be part of this educational process or would it just confuse things? That's a great question, actually. In fact, I think if you look at the experience in Europe, it has become very integrated in terms of the politics and much of the sort of reaction of the European community that has been even more conservative, it has been more conservative than it has been in this country in terms of genetic testing and any genetic manipulation, et cetera, has combined those issues together. I don't know, we didn't talk about that. I personally think that food is probably a lot easier than what we're dealing with. We actually have genetically modified food right now that we're eating. I think a lot of people don't realize that and they're probably not gonna stop eating it if they know. Howard Jacob, Medical College of Wisconsin, I'm sorry to ruin the idea of making millions on a video game. About six or seven years ago, there was one actually that was put out. It was quite nice. You could actually go in and mutate the genes. You could create species. They breed, they do things. So it was all out there. In fact, it was in my lab. And nobody played it. There was no shooting. There was no violence. There was none of this out there. So even in a lab where you think they'd be interested in that, it didn't quite capture the essence. Were they adults playing with it? Or kids? They were adults and well, the kids of course when they were in the lab were also involved in it. But it just didn't have the same appeal. So if you're gonna play the game, they must be bombing the absence. Oh no, Brad had that. Brad had bombing and sex. He said violence is right. So that's an issue for public health. But in seriousness, in terms of going after, I think part of education, and I've been working with some of the high school teachers in Wisconsin, is about changing the way we're teaching biology and the way we're teaching at a level and bringing health in. What they find is when they're discussing health and genetics in the context of their health, retention is much better than talking about slime mold. And it's one of those deals that I think is really changing at a much more baseline level of teaching biology and chemistry through health as another way to get the genetics integrated into this. Because people are more interested in that than some of us like biology, but many don't. And I think we've had some good experiences with some high schools also of integrating the genetics in biology and then raising the LC issues at the same time. And that's a way to send the message early on that you've got to look at the disciplines together. Okay, we're gonna go here, and then two, and then three, and then we're gonna have to stop. Okay. Barbecating from Stanford. I just wanna raise the point, a broader point of asking, within the role of NIH, what should be our responsibilities in the educational arena? And some of the elementary and high school level education, it seems to me, in basic science, which is what it is, I think also we need to look beyond the Genome Institute and look to the Department of Education and NSF and all the other appropriate venues for that. And because I think in a way, it's not a problem we're gonna solve alone. And just to point out to bring back and put back on the table some of the social justice issues just a tiny bit, I've seen, I've been involved. I live in San Francisco. Genentech has a wonderful band that goes around to all the schools, spitting out DNA and doing all kinds of wonderful things. There have been fabulous programs developed by the Genome Institute, which have high tech hookups to high schools. And then I go back, California has a particularly interesting history in public education. I went to my daughter's middle school classroom where there wasn't even a sink in her biology classroom, let alone anything else. And the schools are literally falling apart. So the whole educational infrastructure in science is really the big issue if we're gonna get people to understand these big questions in genomics and biology. Great, great suggestion. I'm just echoing a little bit. I think the issue of educating people, this stem cell debate, I think showed that there was a lot of ignorance out there in some cases. And I think to avoid that, I think it really is a huge job to educate people. And I think it's not just policymakers with the public as well, since policymakers often respond very much to the public. And I think we need to think seriously about funding that kind of effort. I think also getting scientific associations interested, patient groups interested, et cetera, I think needs to be fairly sophisticated and important and urgent. Okay. Helen, last question. Well, it's actually a comment. It's a comment. But since the question was asked, how we're gonna do the next thing, I thought I'd just say a little bit about how we're gonna do the next thing because it's coming, which is how we're thinking about the haplotype map. And I wanna make two points here. One is that having just spent a day and a half thinking about community consultation, that in fact, what we need to do is engage the community and to learn about what their concerns are, to understand their perceptions, to learn what's bothering them, and so we can take those things into account and respond to them appropriately. This is not coming down from Matt Olympus in sort of telling people what's the truth. It's engaging in a collaborative arrangement and that requires conversation. So that's... Can I just respond to that? Our group actually did a lot of talking about that and in the report, and we had some back and forth about do consumers or the public know what they need to know? And I think what came out of it is that even if they don't know what they need to know, they know what they might want to know. And they have views and opinions. So we validate that. It's a good point. I just am saying that the right word should be engagement as opposed to education. The second thing to say is that actually anticipating this project, we're spending a lot of time thinking about thinking about the work and thinking about how we're talking about it and thinking about how we're moving forward so that we can do this in a prospective manner. I mean, we recognize that going out and doing a project where we collect DNA from sample DNA from populations that have arisen all over the world to try to understand genetic diversity and the haplotype runs into some very serious risks. I mean, I'll name them, the Human Genome Diversity Project, racism. And I think that, and we have an important obligation and frankly, the Genome Institute has picked it up to say, we need to identify those and to be ready to think about those in a very prospective manner. So I think that those are two important issues. We need to learn from history. We need to learn from what's already blown up in our face. So we try not to do it again. And then we need to be understanding that this is a process of engagement. And so many of us work with religious communities. Many of us work, because it's not that they're Luddites, it's they're part of our society and they have things to say that we need to respond to. So I would just, I mean, I know that I'm just sort of using slightly different words to say what you intended to say, but I just, but I really, but I wanna put engagement as the word rather than education. Well said. Thank you. Okay. Two announcements. First of all, we will reconvene here at 11.30 for the, promptly at 11.30, because we are gonna run up against time constraints for the wrap up talk by Maynard Olson. And the second thing is I wanna thank the audience for some terrific participation and particularly thank all the breakout session chairs.