 I have a one-year-old and a three-year-old at home, and I have a perpetual cold. So I have a Lawson's and some hot tea. So hopefully we'll get through the next few minutes, but I'm also the last person before break, so I'll try to be expeditious with my remarks. So thank you again, Jean, for including me on the agenda after hearing all the remarks from the various representatives from the different societies. I'm glad to say that NHGRI is leading an effort to develop a resource of clinically relevant genetic variants, and something that we've recognized early on is that in order for this to be successful, we really need the input of the professional societies, clinical organizations and other stakeholders early and often, so we ensure that the end result is useful to the end users, the clinical community. So many of you are familiar with the Nature publication in February 2011, where NHGRI outlined the vision for genomics over the course of the next 10 years or so, and in that paper, there's the Imperatives for Genomics Medicine figure, or box. One of those imperatives is that we need practical systems for clinical genomic informatics, the point that's been alluded to many times throughout the course of the day. We all know that many variants associated with disease risk and treatment response are known, many more will be continued to be discovered. And new models for capturing and displaying these variants and their phenotypic consequences should be developed and incorporated into systems, practical systems that make information available to patients and their healthcare providers. And I think in this, before we get to that latter part, is that we really do, we've heard this today, we need synthesis and evaluation of the available data and evidence to identify clinically relevant variants and make sure that when we produce such a resource, the data is presented in a way that's consumable by EHRs and other point of care decision tools. So as I said in the February 2011 vision statement, this is sort of for the entire genomics community, but as we progressed through the year in 2011, we heard more and more at a variety of meetings, including I believe the first genomic medicine meeting that NHGRI would be well positioned to play a role in helping to bring the communities together to develop such a resource. So with that in mind, we organized a workshop back in December of 2011, so just over a year ago now, characterizing and displaying genetic variants for clinical action. The goal for this workshop and many of you were there, Mark Williams and Rex Chisholm were our co-chairs, was to consider the processes, databases, and other resources needed to identify clinically relevant variants, decide whether they're actionable and what the action should be and then provide this information for consideration for clinical use. There's over a dozen or so recommendations and I'll just highlight one or two, which brings us to the point of the development of this resource. So there really was agreement in the workshop that NHGRI was well positioned to support a resource to extend the existing databases of genotype-phenotype correlation, curating them, adding more information on phenotypes, and coming up with ways to develop clinical utility and actionability. Again, the notion of NHGRI serving as a convener, so bringing together other relevant NIH institutes, professional organizations like all of you around the table and other stakeholders to develop these recommendations regarding clinical relevance, actionability, utility of genetic variants. Another interesting recommendation is relevant here, develop competitions to promote development of algorithms for interpreting genomic variants and compare performance. So it's a little bit different from the conversation we're having today, but again, I think we'd like to develop more automated methods to evaluate these increasing numbers of genetic variants as we move forward. If you're interested in hearing more, listening to the talks or reading the summaries, you can find them at the listed URL. And this is a very simple figure I'm about to show you, but basically we all know you better than I that there's a complex decision-making process when we're trying to decide what genetic variants to use in the clinical setting. We have resources already, genotype-phenotype resources that answer the question, what is related? There's the NHGRI GWAS catalog, OMIM, ClinVar, which is a new database being developed by NCBI, the NIH. There's other resources like ISCA and PharmGKB that also have added value information on top of the genotype-phenotype correlations. But ultimately, what will be done comes down to the clinicians, the institutions, the payers and the patients, and there's a big gap between these two boxes. And so what could be done is, as we talked about, assembling the information on clinically relevant genetic variants and their supporting evidence so that clinicians don't have to reach all the way back to these genotype-phenotype databases and then they can go to a place that has this information assembled in a way that is useful to them. And I just have these arrows here. We've heard Mark ask often, you know, how are you evaluating outcomes? Are we learning about any improvements of using this information? And we just want to recognize that any kind of information we get on outcomes that all needs to be fed back into these different resources. And we want the clinically relevant variants resource or any other similar resources to be nimble enough to be able to incorporate that data and to make sure that the information is updated as new knowledge is being generated. So we have the most accurate information moving forward. So based on the workshop and sort of the thought that we've put into this with many of our colleagues, we developed an RFA. The purpose of the RFA, again, is to support a process for the identification and dissemination of consensus information on genetic variants. And I say process here because we recognize we need to sort of come together and develop a framework, methods that we can apply to this existing data, sort of generate a pipeline so that we can apply it to genetic variants moving forward as we learn more and more. The goals are identified genetic variants with likely implications for clinical care. Establish a process for transferring this information to appropriate clinical organizations for guideline development. Because we know all of you are the ones that make the guidelines that the clinicians use and abide by. So we want to make sure again that we get, we hear from you, what evidence do you need? What do you use to make these recommendations? So we can ensure that as to the extent possible we include that in such a resource, or at least we identify where there are gaps and identify avenues for collecting that information. And then importantly, build upon existing programs, unify, reduce duplicative efforts across research clinical organizations. This is happening in so many places across the country and around the world, which is a good thing, but anything that we can do to bring folks together, to come up with a standard framework or consensus approaches to evaluating this data will be helpful to all of us. So what we're using here for this RFA, the applications have already come in and they're being reviewed. We hope to launch the program in the summer of this year. It's a cooperative agreement mechanism, which just allows NHGRI to work closely with the awardee and other collaborators to ensure we're reaching our goals with this program. We have money set aside through the next four years. Again, we'd like to kick off the program and hopefully summer of this year. And we'll be watching it closely to evaluate the effectiveness of such a resource. We'd also like to see over time, semi-automated and automated methods of sort of curation and annotation being developed so that this process perhaps would get a little bit more affordable and easier to implement over time as sort of the wealth of genomic information is continuing to be developed. So there's some key aspects of the program under synthesis and curation. We're expecting the grantee to survey existing efforts, solicit participation from relevant groups. Again, I've alluded to developing a framework for evaluating the available data and assessing the degree of potential clinical relevance. So you may think of sort of binning different categories of variants in a variety of bins. Could relate to clinical utility and actionability. And then generate a set of genes, variants and supporting evidence for evaluation of clinical relevance to apply this framework to an existing set of genes and variants. And importantly, developing approaches for distributing the curation and the review of variants. One approach I think could be successful and it's sort of evident by the different societies around the room is if we have different specialties come together. So, ophthalmology, GI, rheumatology, all these different specialties, heart and lung, pediatrics, and bring together experts through the clinical lab expertise, the phenotype, the disease knowledge, the medical genetics expertise, have them come together and specialize working groups and then evaluate the data that's been synthesized through the clinically relevant variance resource program to come up with some consensus on sort of what variants bubble to the top and would be most clinically relevant at this point in time. And also identify those that might be not quite there but are close, we have some interesting data, we need a little bit more evidence and try to understand what additional evidence do we need and sort of think about the pathway to collecting that data. Dissemination, again, vetting the specifications early and often with folks like yourselves of the resource. We want this and we heard others in their survey data acknowledge that the users need something that's user-friendly, easy to search, that includes the supporting evidence so there's an understanding of what data was used to make the decisions and then what actions could be taken in the clinic. It's important, obviously, to ensure integration with other data resources, so we were thinking of making sure that the formats we use are interoperable with other existing efforts. And then, again, not only disseminating the information on the genetic variants to the clinical community through this resource but also our approach and our framework. So if other folks want to adopt a similar framework in different settings or in the UK and different places that they may consider using the same approach that we did. And I think the most important part of this in my mind and why I'm happy to be here is, again, ensuring coordination with related efforts. So we have, you know, we know that EGAP has thought a lot about the approaches for evaluating this information. The PGRN, CPIC, they've already done some of this work with pharmacogenomic variants. NHGRI supports other programs like Emerge and Caesar that are generating new pieces of data. They're also thinking about actionable variants that sort of bubble to the surface in their programs. And then, very importantly, the professional societies, all of you. And to that point, I just want to close with, within this RFA, we actually plan to develop a professional organizations committee. So this fits in really nicely with sort of the discussion we'll have after the break. Again, we'd like to bring all of you together as this resource is developed and moved forward to not only give us input on the development and the use of this resource but also to facilitate interactions between you. We'd like to hear more about what kind of evidence, the level of evidence that you need to make these practice guidelines. And we've also heard sort of there's different levels of practice guidelines, maybe not practice guidelines but the next tier underneath that. I forget someone referred to maybe it was recommendations and various other sort of consensus statements. And so, you know, what can we put in this resource that will help you evaluate the data and produce these statements and guidelines more quickly. And as I said, we'll plan to launch the program in the summer. We'd like to gather many of you, those of you that are interested and hopefully I can work with you to identify the point person that would be interested and willing to commit to such an effort. Mostly we'd get together through teleconferences but we'd like to have joint meeting with the steering committee of the program once a year and then, you know, have the representatives from this organization attend the steering committee meeting. So, I'll close with that and just say you'll be hearing from me in the near future as we start to pull this committee together. And I look forward to working with many of you in the near future. Thank you. Great, thank you, Erin. Comments or questions for Erin? Jean. It seems like you're thinking, Erin, about perhaps establishing a society of societies. Is that what's in your mind? Well, yeah, we called the professional organizations committee but again, we just like to bring many of you together. We recognize everyone's busy but we really think that it's critical to get input from all of you as soon as possible to make sure that we're thinking in the right, you know, we're in the right frame of mind as we're developing this resource. Erin, I note that the Wellcome Trust co-sponsored the original workshop we had. Do we know what they or other funding agencies are doing in this area, either based on that workshop or independently? We've had, so there is a manuscript that should be coming out from this workshop any day now. And so through the development of that manuscript, I've had some conversation with Tim Hubbard and other folks at the Wellcome but from what I understand at this point, they haven't invested any funds in this area but they're certainly interested in what we're doing moving forward. So I can just follow up on that question. I was just in England yesterday before with Tony Brooks and he was commenting that there is initiative related to Healthy 2020 in Europe and with interest in funding activities in this area and Tony is in the process trying to propose a meeting in April in Ireland related to variant assessment, pathogenicity, organizing gene resources. So there does seem to be some interest in funding activities in this area but I think there's still not clarity as of yet. Emily. Thanks for that talk. Of the resources we need, I mean, one of the things we need is just more data on what these variants do. And I see a database for variants that are juried is a component of this collecting more evidence that allows us to jury these variants? Yeah, I mean, we recognize that's really an important aspect of this entire pipeline and there are efforts, so I've talked about ClinVar and there are efforts in the research community to try to pull together various labs and other research groups to deposit information on what they're seeing in the clinic as they're applying clinical sequencing to establish, you know, to build up that knowledge base so then we can sort of figure out what bubbles to the surface as we let them. Terry. Yeah, I think Gail makes a good point in terms of the need for additional evidence which we recognize. We wouldn't do that probably as an integral part of this initiative which has a specific focus and a target but I think one of the things that we would like to hear from that resource as it develops as we would like to hear from the professional societies is where, you know, what kinds of evidence are needed, what would be convincing to them. Where are points that are almost there and if we only had this piece of information we would know one way or the other so that's what we're hoping for. And it gets back to the business of randomized clinical trials or something proximal to it, doesn't it? Really. So Bob. I think on that topic I realize that, you know, thank you very much here and I think that that initiative you have certainly and what we're all about today is to better educate and how to get it out to the clinical application and to look at those clinical applications. But certainly I know that the Wellcome Trust is interested in collecting the variants for multiple diseases and what they are hoping to do and I'm hoping but I would like to bring this point to because as I indicated for coronary artery disease it's similar for diabetes, it's similar for many societies. All these variants, most of them, are not acting through the conventional risk factors and yet it's hard to believe that they're all acting through a different mechanism. They must be acting through two, three or four or five different systems and I think that again this is a very complicated area it needs lots of bioinformatics that I think is going to be done individually and will take decades. On the other hand this is the kind of project that the people who collect all the variants and have all of that data somewhere along the line it's easier to put together some framework that would be very helpful for us in terms of individual laboratories to do. I hope that would be the case. Yeah and I think this is really gonna be important to kind of set a community standard here because we've certainly heard about in the context of the discussions around patents and the fact that patents may ultimately not be upheld or will expire related to human genes. There have been some that have expressed that they will deal with this issue by imposing trade secrets on interpretive information associated with variants and that's the business model going forward. So that would be disastrous I think for the group as a whole because trade secrets of course do not expire as Coke can attest. So I think that if we can really set a community standard of collaboration and communication of results and that it would be hard to hold against that type of an effort. And I just want to say that's why I try to emphasize that the keywords process and framework. I mean ultimately the end result will be this database but just as important is developing this sort of pipeline these processes so we're all sort of approaching it the same way. Now Mary are you sure you have a question now? Sure. So can you help me understand? I feel like I've asked this question a couple times before but here it goes again. How compare and contrasts ClinVar to CRVR? So the way that I'm envisioning and we have been talking quite a bit with our colleagues at NCBI they're developing ClinVar. ClinVar and if you think about sort of the genotype phenotype correlation ClinVar is serving as the repository of an archive data repository of genetic variants and then the sort of the correlation with a variety of phenotypes. What we see I don't envision that they're gonna have sort of the evidence on clinical utility sort of other pieces of data that one might evaluate for deciding if something is clinically reaches the standard of clinical utility or not. So we see ClinVar as sort of one of our primary data points. So as clinical labs, from my understanding, are considering depositing information directly into ClinVar so we'll have an increased number, increased information on variants with a variety of phenotypes but then our resource and our process will come in and use ClinVar as one of the primary pieces of data, figure out what other piece of evidence do we need that's not in ClinVar to decide if something is clinically actionable or reaches that point of clinical utility and then that information, so there's that top tier of variants would be displayed through this database. So you will have further clinical utility information above and beyond what's planned to be in ClinVar which will explain any associations between variants be they clinical or otherwise. Right, and they'll envision variants of unknown significance. They would like to think about the entire genome so it's much information that they can have on each variant in the genome and you can envision that's probably not gonna, it'll be broad but not perhaps very, I'm not using the right words. They just wanna make sure that every variant has some piece of information associated with it but they won't have sort of the curation level efforts to pull in the sort of the other pieces of data that will need to be used to decide on clinical utility. And plus the CRVR will have a consensus process developed and applied to the variant ticket so there not only be the curation but a group of experts for one of a better term that will review that information much like. Much more selective. Yes. Yes, it's definitely much more like the CPIC approach. Other comments, yes. One last comment, I mean I think also certainly from speaking for the polygenic disorders it's good to keep in mind that all we have is a SNP in a region in 90% of the cases. And so if you collect data from different disciplines, stroke, heart disease, diabetes, I think there's a lot to be gained by because we are claiming different SNPs when we're all sometimes only within a few thousand base pairs of each other. It associates with diabetes that part of the molecule. This one associates, it may well be the same molecule or altered in place form or different domains the same. That will be much easier if you're collecting multiple phenotypes with the SNPs to put it together which I think none of us who are all interested in our own disease would like to do. Excellent point. Last comment, Jonas. I'm just trying to understand CRVR framework. So is this something like the TCGA but without selecting the population that is being sampled? Say that again something like the what? The cancer genome metals in the sense that you have multiple phenotypic instances of variances that may be the same. Well I think in this regard and the reason I put up sort of the different phenotypes or disease categories is just that we recognize there are examples of some variants that act in different disease pathways but the approach that we plan to take was really just sort of thinking about disseminating curation efforts so we'd like to sort of pull together the disease experts together with the medical geneticists to evaluate the evidence at hand and then make decisions on the clinical relevance of variants sort of in each of those disease settings. And I could add that the cancer genome atlas is actually generating the evidence so they're doing the sequencing, they're identifying the variants, they're looking at the phenotypes. This is really focused on taking what's existing out there so what existing literature and information is available and then putting a consensus process on it with the hope that there will be continued generation of new evidences as time goes by. So we've generated some further comments maybe we'll take two more. If we had John and someone else who others, who else's hand did I see? I'm curious if you are considering linkages disequilibrium as a curation as in terms of imputation or using the markers that might be done and what they might mean in their neighborhood. I think that's a valid comment and I think when we get down to that level of detail it's sort of up to the grantee and the approach that they wanna take sort of to assemble all this information and make decisions on how we're gonna evaluate it moving forward. But it's a good thought to keep in mind. I mean we're not in a position right now to design the project because we're in that sensitive period when applications are in and being reviewed but a good thought for us to consider. Thank you. Anything else? I thought you'd give us a plate. Oh yeah, sorry. Sorry about that. All right, so I think we're up to a break. We're gonna take a full 30 minute break so that we give Gene and Mark a little bit of time to put their efforts together for our next discussion. So we'll come back at 20 minutes to four and we'll give you a two minute warning. Thank you.