 So, I think we need to go on, but we're scheduled to have what we called an LC pause, and I think it will be obvious that the few remarks that I want to make underscore the relevance of this conversation because I think there are rules of the road that are going to have to be determined and they have to be informed by science but also by considerations of policy. So, let me just start by saying we actually are acquiring a growing body of empiric data about what participants think and what participants want around the issue of genomic data and around the issue of sharing of genomic data. And I think there are some very, very consistent messages emerging. Participants in research by definition, people who have agreed to be part of the research process are strong supporters of research in general. They want research to be pursued and they want it to be directed toward benefit. And when you probe, and I think where our data are more limited at this point, but when you probe, it's pretty clear that there's a strong endorsement of societal benefit. And what I would say is a buying of the message of NIH. That is a buying of the idea that biomedical research leads to health benefit. So participants want to see that health benefit. They're very supportive of that. They value information about how their samples are used and that is separable from their interest in the research being pursued to generate benefit. They're interested in knowing how their samples are used. We've certainly heard from participants a great deal of pride in learning about how their research is, how their samples help research move forward. Interestingly also, I think in our studies and other studies, there seems to be a consistent expectation that research participants will be told when there are research results that are relevant to them and that can provide personal benefit. And here we see that among research participants there is perhaps an unrealistic expectation and an expectation that gets more unrealistic the more we build the kind of data resources that we're talking about at this meeting. Which I think has implications for how we talk to participants and what kinds of promises and what kinds of transparency we provide. I will say it's an untested issue when you look at bullet two and bullet three here to what extent participants might be very satisfied with more information about aggregate results of research because that isn't very often fed back to participants in venues or language that is accessible to them and probably is something that we need to think more seriously about. I'm also including in parentheses here certainly for the U.S. it's important to acknowledge that there are substantial numbers of people within our population who are quite mistrustful of research. We have data to suggest that's higher. Those numbers are higher in individuals from minority populations. We also see from the controversies around newborn screening that it's certainly by no means exclusive to minority populations. There's simply a subset of individuals who are very mistrustful of research and we need to acknowledge that. I think it simply underscores our need to be responsible in how we communicate with participants about the research process. I would propose as something we have to seriously think about as we think about the data resources here at this meeting that there is a limited what I would call a limited ethical reach to the global consent. Informed consent is intended to support autonomy. If you go back to the Belmont report and see what we're trying to accomplish with informed consent, it's to allow the voluntary participation in a research process that the participant is informed about fully enough that the participant can determine the risks and benefits of the process. That's not what we're doing with global consent. I think there are all sorts of reasons why first of all creating data repositories and sharing data produces research benefit. I think there's a logic to why we would use a global consent process, but what I think that tells us is that we can't stop with informed consent process. That's part of a responsible research enterprise, but it's just the beginning. Participants do have an interest in prospective engagement. We have to acknowledge the return of results issue. As I've mentioned, there's an expectation that of course researchers are going to tell me stuff that's important to me. We have to be honest about when we can and when we can't. I think we have to be very clear about why we can't. We have to be very comfortable with the idea that we may be creating research repositories where it really isn't practical to expect individual return of results. The weighing of resources going into that process versus going into research may not make sense, but we need to engage participants in the conversation about that. We need to also think about the fact that even in the global consent, participants retain the right to withdraw from research. That's kind of a bedrock of autonomy protections in the informed consent process, but to meaningfully withdraw from your sample from DBGAP or a central repository, you have to know what's going on. A meaningful withdrawal opportunity means that we need to think very carefully about how we inform participants about how their samples are being used. In fact, that's something participants want us to inform them about. They want to understand the research process and they want to understand what outcomes are available. All of this, I think, leads in a broad way to a need to think about stewardship of data resources as much as technical details as we talk about what would be involved in streamlining DBGAP access, central repository, or any of the other models that we're considering. I want to pause here and just give you a little snippet of data from one of the focus groups that we've done. This is in an urban HMO, a predominantly white, very well-educated population, very supportive of research, a source of data for the points that I've just been making, but they're also quite mindful about realities in terms of data protection. There was this little back and forth between two participants in a focus group talking about you create a database and sooner or later there's a computer that gets lost, somebody steals them, somebody hacks them, whatever. This happens. We see this in the news it was referred to earlier today and the second speaker comes right in and says, however, having said that, we don't stop using banks. I think the point here and the reason why I'm showing you this particular quote is that I think participants in research, by and large, are going to be realistic. I don't think we need to make unrealistic promises to them, but we need to make promises that are as accurate as we can make them. We need to figure out what we can promise and what we can promise, and then to go back to a discussion from an earlier session, we need to make the case that participating research is a good thing and generates benefits and so to the extent that there are tradeoffs, there's a reason for those tradeoffs. I would propose and I just have three more slides that there are three elements that are key and we can think about what all the different details under those are. One of them is governments. We really have to ask ourselves what model or models, there may not be a single one, of governance are going to be widely accepted as reasonable and fair as we build these structures, and that means we have to think about who's at the table making decisions, what protections can we promise, how is decision making authority allocated, how are differences resolved, all of these issues are going to be on the table. Any time you have a governance of a structure and we need to go into any kind of data repository structure thinking about these things. Transparency I would argue is a hugely important component and one that is largely lacking now in the interface between participants and genomic research processes. That means transparency about what the governance procedures are, who makes the decisions, if somebody is going to be certified, who does it, under what criteria, how does that happen, exactly what are the procedures for data protection and what's our best estimate for what they provide. And then transparency about the research process, where did the data come from? I think it's a reasonable surmise that a lot of the people whose data are in DBGAP don't know it, it wouldn't necessarily find it easy to figure it out if they do, and I think over time that's likely to be a problem for us. I think we need to make that information easy to find, I think how the data are used need to be very clear to people and what comes out of it. Most fundamentally how can we back the argument that the creation of these data resources, these data sharing opportunities do accelerate the process and get us quicker to the kinds of benefits that participants care about. I think we really have to figure out how to explain that because lacking that sustaining participant interest long term may be a problem. And then finally accountability, this has come up already and I just want to underscore it, things will go wrong. So what happens when they go wrong? And how do we know that they went wrong? What are the criteria for certification and who's checking that those criteria are really being met? I think we can guarantee that as we create a certification process there will be people who submit fraudulent data seeking certification. So how do we set up a system that's not onerous but is reasonable? How do we keep track of whether data are used according to the rules that we've all agreed are fair and reasonable? And most significantly what kind of consequences are there for bad behavior? I think the point about I know someone will go to jail if they do something wrong with the money in my bank is relevant here. I think part of generating trust is making sure that the consequences are there. So that's all that I have to say. I just want to say I think as I'm listening to the meeting and our discussion is going forward I think these are issues that need to be out there and need to be part of our discussion as we resolve scientific issues around the data sharing challenge. And Mike, I don't know how much time you want to take for discussion now. David? So I guess I mean I think everything you said is very reasonable but I guess the one thing that I'm puzzled by is I think you're just in terms of saying we need to explain to people why this would accelerate progress. I guess what you're saying is we need to disabuse them of the notion we must have created that by sequencing a genome we could learn anything at all because in my view sequencing a single genome the only way we can learn anything is if someone previously sequenced a lot of genomes and found a pattern that there was mutations in BRCA1 and looked carefully at people who carried them and figured out what was going on and for anything we don't already know about it's a comparative process and I know some people may not agree with that but I think they're wrong. So you know so in other words how are we like what you're saying I think is we have to disabuse them of a notion we must have created that somehow in the absence of comparative data we could interpret all this stuff because I can't imagine how else otherwise they wouldn't understand. No I think there is some hype out there and I think there is some room for disabusing people who have simple and easy things might be but I actually think that what is actually going to happen when we take 70,000 genomes or 700,000 genomes and what kind of data are going to be important in non-genomic as well as genomic data in making sense of that and actually slowly figuring out you know biomedical problems and what we might do about them I think that process is largely very opaque to much of the general public. I think we have an extraordinary opportunity to explain that because we have an extraordinary opportunity to make progress and it's going to take resources you know it's more than us having a press release about nifty papers it's going to be probably gathering data about what kind of messaging really makes sense to people and what pieces of that process are most opaque and how we can explain them better. Debbie. I've resonated with that in a very different way rather than overselling I think just telling people the people who consented to be a part of any of these studies that go in are extremely altruistic not knowing what was going to happen they did it anyway even if you wrote down everything that could potentially happen to them they did it anyway so I think that actually just laying out all the positives that could potentially come from this not promising but why a resource like this is very beneficial also helps and I think I think calling for that is really very important I agree that it's not to explain things have been over promised it's how this resource can generate new insights into human biology and science and it's not necessarily gonna gonna generate tons of new papers or tons of new insights it'll be a gradual thing as a data increases but these are the kinds of tangible things that we've seen and you could make models for how we've seen this as genomic has proceeded and how we've learned more and more and you could make that available on on the site or something. I think that's a great idea I would expand it though and try to avoid genetic exceptionalism and really go all of research because unfortunately we're finding is people say well I said no the genetic research and then it's like I said no to all research one thing we are exploring is research notification to do in a very celebratory manner this is how you know these are the types of research we do and congratulations and we're having a hell of a time getting it through our leadership because you're afraid everybody's gonna opt out it's just it's been a fascinating it's been a four-year navigation. I think as productive and creative as this whole enterprise has been with sequencing and the transition and development technology over the years what we've failed to do is even come close to parallel that with an ability to converse and communicate with research participants you know and Pearl showed the language they're using now and to think that this is what we're presenting to folks in a written form and expecting them to have any real concept of what this is because this type of research in particular I think is mostly conceptual you know it's not like telling somebody you'll take a drug and you'll probably get sick and your hair will fall out I mean this is all very abstract sort of information so I think there's real opportunities to try to think more creatively about how to communicate with animation and video and film perhaps different technologies that can be brought to bear in this domain to give people a much better idea of what we're talking about with this because I certainly agree with you I think the whole notion of having ongoing communication with research participants isn't likely to be successful so we have to up front get their trust and allow them to say yeah this looks good to me I trust you do what you think is best with this sample to promote science and then live up to that trust. Pearl did you want to one sentence I think we also need to focus us on health care providers who can be some of our best teachers. Good point. Any other comments? Oh sorry last comment. So I want to raise a question really back to your accountability slide you have up here right now some day they'll be a posting by WikiLeaks of a big chunk of DB Gap somewhere on a bit to our insight where are we going to do then and what will be the restrictions on people in the biomedical research community on using those data which they can now get for free and anyone else can get for free from that bit to our insight. From that what site? Where they can download some data that someone stole from DB Gap and posted it in the same way WikiLeaks posted all the Department of State cables. So just we already had a leak like that a couple years ago with AOL's data so you know all of AOL put the de-identified search queries of 650,000 consumers back in 2006 that was re-identified by a couple New York Times journalists. That wasn't really a leak. No it wasn't a leak it was an explicit disclosure but you can still get that data set on torrent sites and it's basically taboo you cannot publish on that data set and anybody who has tried to publish on that data set has basically been barred from the information retrieval research community. Well I think that point that speaks to two things one is we have to be realistic. We have to be in conversation with the public about what we're doing and why we're doing it and potentially what risks and if we think that there is a realistic possibility of a scenario like the one you describe we have to have consequences in place. I'm not trying to say any of that is simple but I think the alternative approach which is to not think about it not talk about it and then have it be a big disaster when it happens it would be a big disaster for the research process so I think that's why we have to be prospective about it.