 Xander, are we ready? All right, I'm gonna go ahead and get us started on this last but in our opinion, most important of sessions. So for those of you who don't know me, I'm Jennifer Troyer. I am the incoming director of Extramural Operations for NHGRI and so speaking for all of us here, I wanna thank you all for sharing these last couple of days with us. It's been exciting talks, exciting discussions, great panelists, great questions. We thank you all from coming from your various places in different time zones and bringing your expertise here to think about complex traits. And now we want your collective thoughts on what is next. So we've talked about all these different scales. We've made very clear that LOIC is going to be working for a long time. I'm getting from the genomics to the trait, to the environment, and so are all the rest of us. And so what we really want to get from you keeping in mind that NHGRI is listening, but so is NSF, so are many of the other ICs at NIH. So are I here up to several hundred people online and around the world. So your advice can be to us, but it can also be to others and to the community. And we want to know several things from you. Everything from what is the moonshot? What is the, if we could do whatever we wanted with infinite resources, infinite data, what's needed, but also the practically, if we have limited resources, limited time, where are those leverage points? What can be done now with small amounts of focus and investment and bringing people together to move the needle in large amounts? And so all I ask from you is when you make your suggestion, if you could say who it's for, is it for everybody? Is it for NHGRI specifically? Whether it is moonshot, whether it is long-term, whether it is short-term, whether it is practical, whether it needs money, whether it needs people thinking about it, and then to what end? What will we have out of it if we do it? So bring your ideas, bring your focus to give your recommendations, like I said, to this multi-layer of people listening about the multiple things that we could be doing in this space to improve our ability to gain knowledge from genomic and other data. And so there will be four microphones running around. Those of you online, somebody is watching the chat, we won't necessarily get to all your suggestions, but we ask you to put all your suggestions in Q&A. Somebody will read them, I promise. And if nobody else will start, oh, we got one, all right. So I'm gonna make a suggestion about which I don't know very much, but I think you could invest in filling in Tully's Valley of Death in between the molecular data and the final phenotype of disease, for instance, all the biochemical and physiological steps in between, and this would be a very useful database for everybody. So it's not a sort of a solve diabetes type project, it's an underlying database for everybody, but there's an important design question of whether you do all the different things on different people or the same people. And I'm not aware of any active design work to compare those two, but I'm gonna suggest that you do them on the same people. I think there are advantages in a design like that because that's not what we have at the moment. We have designs where some people have done the expression QTL and other people have been used for the disease phenotype. If you had everybody done on everything, then you could estimate environmental connections as well as genetic connections. Okay, next, back here. So this is a suggestion that I think is tenable, but it would require data and investment in transdisciplinary groups. So I've heard a lot of talk about context, but it's almost all been described at the individual level. And then we did talk about family and populations, societies, geopolitical, but I think an investment, an intentional investment in data sets that would be diverse but have multiple levels of environmental influences is important because what we might see is that it's really structural. It's structural racism, for example, it could be. It could be the brain drain, but what we're gonna see at the individual level is just a reflection of that. So we might be looking in the wrong place. So Aravinda started by saying, why is it that people with many different combinations of risk variants have the same disease? And so I think that's a question we could address. And so Nancy suggested doing high and low polygenic risk score individuals. And I think Jonathan talked about doing cell-based assays. So the design I like is doing cell-based assays for high and low polygenic risk score for people who are cases and controls and then generate cellular phenotypes that multiple, do a high-scale, multiple cellular phenotypes that you can read off in many ways. And try and find combinations of measures that would show us why people with different risk or variants converge to the same biology. I just wanna compliment that. I think that's a great suggestion. I was wondering, I guess I'm trying to make that point about it this morning, whether height will be a great place to do that because I'm thinking if you contrast the top, the tail of the polygenic score, we're talking 24 centimeters difference and at least at this point in time that's a massive effect. And maybe that's a great place to start for this kind of experiment. And the cell types could be, you know, condro sites or anything actually, but yeah, just a compliment. Oh, thank you. Just a quick comment. So the idea of using polygenic risk score to stratify, it doesn't seem to be optimal for the following reason that there are many different cells and many different biological phenomena into highly polygenic complex trait. So if you just go by polygenic risk score, you can influence them. So if we can select endofinotypes and pathways and specific components, like if you do cholesterol for heart disease or other situations like that, lower level, and we can tease them apart from the data, it might be better than to use polygenic risk score as a unified measure of something very complex and very multi-dimensional. I wanted to highlight some of the things that I heard that I think we need to agree on in order to suggest the best possible experiments. And it came up over and over again. What are our key questions? And I think there was a great articulation in one of the sessions today. We care about mechanisms. And for mechanisms, there's no question we need more biological data to be generated. There are many ways to think about how we need to have that data generated. It can be at the cellular level, it can be at multiple levels. And I think those are things that we need to spend some time thinking about. But I think we can all agree on that goal. I think another goal is clearly some better understanding of context and the ways that it's going to inform us about architecture. And so that's a great goal, but it's an even more complex question from the design perspective. And maybe we need some pilot studies to inform us more about the levels of context we need to look at, how specific or nonspecific contexts are. And I mean that sincerely because lots of drugs have similar effects on not just gene expression phenotypes, but methylation on chromatin confirmation. And so you can have similar outcomes, similar biological consequences of different contexts. And maybe that's another part to this that we need to learn. But I think we need to agree on prioritizing some goals and then designing the best experiments we can to achieve those goals. But also we used to do so much more modeling. We used to depend on that modeling for how we thought about what data we needed and how to design the experiments. And I do think that we have more understanding now of the complexities that ought to be, but are not yet in our models. And there needs to be some investment, I think, in model building around them. So that's all I'm gonna say for now. I just wanted to second what Nancy was saying because I think at the end of the day, most of what the people in this room are doing is based on models that were developed a century ago. And clearly we've gotten a lot of mileage from it, but we've also heard all the ways in which these models are starting to fail us or have already failed us. For example, in considering that humans are unusual in the fact that we inherit much of our environment as well as our genetics distinct from many model organisms or in thinking about why it is that traits are so polygenic and what the omnigenic model and potential extensions of it have taught us. And also in thinking about where environmental perturbations come in and whether we should expect them to come in in the same low sigh as genetic perturbations, all of these things require new models. I have a mic so I'm gonna go. One theme that's jumped out to me throughout the sessions, going back to Aravinda's talk and his example of her spring disease where you have a big gene regulatory network and at the end there's these two different genes expressions of which matter. I think and also in NASA's talk, we had this notion of multiple different kinds of causal factors being considered together with the idea that if you look at them together, you're going to identify some kind of higher order organization in which they sit. And I think if going back to Jonathan's breakdown of goals, I think if your goal is prediction, maybe you're not so interested in things like epistasis because there's not a lot of epistatic variants, but if your goal is mechanism, then finding a case of epistasis is really useful because it tells you that these two things are connected in some kind of system. And I think maybe we haven't looked for epistasis in particular or other kinds of interactions because it's very hard and because the, you know, the amount of variance explained is gonna be low but I think if we're interested in mechanism, it might actually be a very productive thing to look at. I saw another hand in the back. I wanted to second the thoughts before about looking at more models and a better understanding of context. And I think one of the best ways that we can do that is to ask people who are affected by the traits and diseases that we're considering and better centering the work that we're doing in the community. I think there's a lot of work that's being done that's like focused on building a system and like knowing more about the context that we're working within and knowing that people who are affected by it more is a really helpful way of developing a more holistic approach to this that we can use in further studies to better understand how these factors are related to one another. Just a follow-up on Molly's comment, actually thinking about what happened the last time we had this 20 years ago, the big concern there was our models were outdated and don't work. But the one model that's consistently held true has been the mixed model because it's so flexible. And so the thing I wanna suggest is we certainly need not necessarily new modeling tools because we're out there. We just need to look at them in different ways. There's a huge amount of literature in agriculture and an evolutionary quantitative genetics that can address some of the questions here. Like how do you handle the situation where your environment itself changes because it's inherited? Well, that theory is well worked out for a lot of things in animal breeding as Mike alluded to. It's a lot of rich machinery that I think can be exploited. So I think it's important we think in different directions but we all have a tendency to be very siloed in quantitative genetics. And there's lots of independent machinery developed in other fields that's really quite relevant. You have to tailor it, but it's sitting there ready to go. It just needs to be tuned up a little bit. That means no funding, so I shouldn't probably bring that up. No, that's awesome. Get back. I would like to see integration of history. It might be nice to see a specific RFA of how history has impacted population variation or the expression of certain genes. I'd also like to echo the thoughts on community engagement. It is possible to have studies. We do this in public health where localized communities own the genetic data and then they enter a cooperative agreement with you to use the data to look at whatever you sort of want to study. So I think a funding announcement to allow interested investigators, probably early career investigators, want to do this to build a relationship with a community and then extract data from them in order to report back to the community what they see. So one thing I know about giving advice is most people don't take it and that includes me. So I'm gonna be trying to be a little general. I'm glad to hear we're just like everybody else. And I found Jonathan's articulation. I was trying to get there, but he articulated basically the three classes of problems, if you will, in which the aims are to learn biology. I sort of like the fact that you use biology and didn't say a particular kind of biology. Mechanism means different things to different audiences and communities. The second is to improve and personalize risk estimate and make it far more general and accurate and there is other applications of genetics which may be extremely important, particularly society's views of what are other studies in one and two need. So I'd like to make an appeal for really four kinds of things without going into the specific questions. One is again, I think Nancy and Molly spoke about models. I think there's enough data out there for people to propose, however they analyze these models, to understand the kind of data that may be necessary. And that's as a first step. I really would hope that, I think we've gone through a success period in genomics over the past. God knows how. It's large community projects, but I think solving the problem of complex disease and specific questions may be better answered in one system than another, rather forcing a whole bunch of investigators as often happens into one paradigm. And the third related to it is, just as it's useful to have a broad set of phenotypes for people to test their ideas, it may also benefit to have a few exemplars, like maybe one of them, to test all of the ideas, to understand how far we may be able to go. We don't know what the stops are. Maybe we won't get the phenotype very easily. I'm sort of in doing, for example, model systems, meaning usual model system or model organism should be part of the mix. And the last is really, I'm pleading that you give individual investigators the option to ask questions and fund it through R01 mechanisms, which is rather than only large consortium grants. I was gonna pick up on what NASA and AJ both mentioned in that when we think about an institution like the NIH, which is funded through taxpayer dollars, members of the public, we need to consider what the downstream implications of this research is for those people that are making the work that is being done in this room possible. So I just really wanna emphasize that I think if we're wanting to, and I think we need to think holistically about the risks and potential benefits of this research, which is critical when we're utilizing a scarce resource like federally funded money, taxpayer dollars, that we have to think about how the research that's happening upstream is affecting the people downstream, who again are making it all possible. I wanted to echo some of the considerations that were already being shared and specifically focused on this idea of diversifying. I think we can think about it as not only diversifying by who we sample, because that's certainly something we want to do, and by diversifying who we sample, we're also going to diversify which environment we sample, as long as we actually measure as well. But we also need to diversify the biological level that we're looking at. And so, I think if we bring these two aspects together, we can ask questions both in terms of translating genetics to function, but also making better predictions. I want Loic to pitch Loic Biobank. I don't want to steal his idea, so he should pitch it. I think it's a really good idea and relevant to what we're discussing. Okay, well, I guess that's... Yeah, well, we had a few drinks last night. So, well, the point was not to name it the Loic Biobank, but the idea is to essentially think about ways we could generate new ideas about things we don't know that we don't know. And I was thinking about the analogy of adversarial neural networks where essentially you have one network that is trying to produce an outcome and the other one is trying to tell whether it's a true thing or not. And I was thinking in terms, as a community, we could think about using our best models about what we think complex traits are. We've learned about new functional annotations, where the heritability is supposed to be, et cetera, et cetera. We can come up with our best models. And the idea is to say, can we come up with a neural network that can tell, is this simulated trait that you, which essentially summarized everything that as a community you know, is this simulated trait different from a real trait that you would have GWAST? And I was hoping that if we, being creative, maybe we'll see, I guess the two outcomes, I won't be too long, two outcomes from this. One is either we will validate that some of the assumptions on which we're building our models are actually very good. And that will be if the models cannot tell the difference between a real trait and those simulated traits, simulated GWASs, or if we can find ways to learn what are the features that this neural network are learning that we haven't encapsulated, we could essentially generate new testable hypothesis. So just a specific suggestion of theory that I think is useful to be developed is we've talked around a little bit, and that's the whole idea about two emergent features arise as you change levels. And... As you change microphones, new ideas arise. Okay, so, such a... Yeah, we should probably charge one of those while the other two are being passed. Yeah, that's a good idea. I think we can also pass these. Okay, we'll try this once. This is what's called the G by E. When you go from individual cells to organs, but also when you go from individuals interacting communities, a lot of the mathematical concepts are relatively similar. I mean, the details are different, but in terms of putting them in terms of graph theory. The other really potentially interesting thing that arises is that when you get nested levels of complexity, are there emergent features that fall out just like if you add a bunch of things that they've roughly followed a normal distribution? Are there features that arise at that level, such as how come many biological systems are relatively stable? Is that a highly evolved feature? Is that just a feature of how things are put together? So I think that's an interesting level of theory that kind of goes over multiple levels. And I think you could get some potentially interesting experimental design to test that as well. All right, so while we're waiting for the next hand and the microphones to charge, I wanna push a little bit on a couple of the ideas that people have expressed. So one is in this area of let people just try things, RO1s, right? Which everybody can always submit your RO1s to the parent. I think right now we have some RO1 announcements and R21 focused on this area specifically and really focused on modeling. But what are the things that would support the ability of individual investigators to do their best science in that RO1 realm? Are there resources that we need? Are there ways of looking at tools that are needed, data that's needed? So are there things that can support that RO1 work? And then I wanna hear more about the, we've talked some about involving the community on the front end. We've talked some about making sure the community benefits on the back end. But what are specific things that would go into making that happen well? I got you, no problem. So I have one, sorry. Okay, sorry. So the way that the mechanism might look is, and I know that we're stuck with the five years but I just wanna push and say sometimes seven or eight might be strategic and that's why I said early career because it might be some other kind of mechanism where you allow people to spend two years building a relationship with a community partner or organization and then they spend year three and four collecting data and year five and a half, six determining what that data actually means for the population. So that's sort of how that can look. What it would also need to include is funds available to fund the community partner. And in a probably a better equitable way would be an MPI situation where the community partner actually comes in as an eco-partner to the scientist and has their own byline item in the budget so that you then don't have to worry about the scientific organization taking its pretty time to send money to the community org to help out with identifying participants who might be a part of this study. It also looks like having a dissemination budget that might include storage, community partners. So some tribes actually have anthropological partners where they store the data on their grounds and scientists can come and analyze the data on their grounds but that stays there. So dissemination stuff might look like storage facilities it might look like training for community members to handle their own data. And those are sort of some of the approaches that it might look like in a grant announcement. Magnus. So I actually have a very simple point it's something that probably contradicts what was just said but as someone who comes from outside human genetics a big something broke is simply the availability of data. I'm used to all data phenotypes and genotypes just being available, right? I realize this is a very hard problem to solve in practice but it actually makes a difference for modeling and analyzing things. It sort of democratizes things if data are available. Yeah, to Ajay's point, I just put a grant into an NIEHS call for climate change modeling and that require community engagement core and combat that with the social scientists at Georgia Tech that I was working with on that actually building the community structure was really interesting for me and we'll be fascinating going forward but all the elements you just said were in that proposal mechanism. So that's there. The other thing I wanted to I guess argue for would be I mean, if you thought about a next ENCODE project would enhance your eye. I mean, variant to function code project. I mean, a call for RO1s to have people strategize about using CRISPR-i or base editing or whatever in primary cell lines in different contexts to really map out the function of GWAS hits for the relevant disease. I think that could be a vision that sort of took us from the GWAS to the function. But a nightly at the level of our one investigation. Second expanding upon what AJ said, I think there's a lot of space here for creating a different structure for financing the science that we do. And the options there are kind of limited by the structure that we currently have for like how grants are submitted and the structure of the way that this work is done. All we can do is have people as MPIs because we have to have a PI who is like sponsored by an institution that creates all this space. And I'm wondering if there's a way to like have a funding mechanism that is basically funding organizations to provide the grant infrastructure and provide the training to allow community organizations to submit their own grants or to have funding available so that community organizations can directly access money to work on the research that they want to with the support of scientists rather than having scientists work on the work that they want to with the support of the community. On that one, it may not be exactly what you want but I don't 100% mean like we've solved it but there's actually a really interesting program at the Common Fund right now called the Compass Awards which are community led projects where they're actually are trying to come in and do the funding. It's been very interesting whenever we bring people who aren't typically funded by NIH into NIH there's a lot of obstacles. And so part of what that program is also doing is helping work out what are the sort of challenges to expand who we fund into those types of places in ways that would allow such for future things. So I don't mean, oh, there's one program that's all been solved but I just wanted to call that out because it's been a really interesting experiment. It actually broke some of our systems last year and like OER had to get involved in stuff but I do hear what you're saying and I think it's not just reaching out to those groups but understanding how to fix the ways that you can get NIH funding that isn't just, I mean it's that isn't just so much of how we fund is based on the way that and I U.S. academic organizations are structured. Can you hear me again? I don't like the people with malpheonomics initiative as well. Yeah. It's not a perfect compliment but it is data that's relevant for people in the room for some of the questions that we. Yeah. So we have two like projects that I wanna mention the impact of genomic variation on function which is doesn't have the data available yet but that we've started which is trying to and you know, people are involved in it. I'm involved but I'm gonna people in the room who are involved might describe it a little bit better than I do but I think that the idea there is to look in different ways and some of it is falling up on GWAS hits and looking from a CRISPR perturbations and developing models, et cetera to look at can we look at molecular cellular function consequences of variance and develop a catalog in a way that's gonna help some understand some of these generalizable principles and sort of move some of those forward and there are some pilot projects there that are looking in some of the ways that were just described not for our ones. There's also our ones that we fund in that but from a large scale project and setting up also a database that will and a catalog that could be built in and used. Nancy was mentioning our more recent, I saw, was that a hand up? Go ahead. Okay, thanks. I just wanted to comment about IGBF and our experience with that and to relate it to what Tullis Valley of Death and what Michael Goder just mentioned. So IGBF is variant centric activity with the mapping awards, the perturbation awards. It's around a variant, right? And the discussion here, what I heard about cells and what I heard about is about phenotypes around the phenotype. So the example of a project that is very difficult to run through UDN is I have a gene of unknown function. I think it's important for my phenotype or it's important for natural selection in humans. I have a pathway. I want to find a proxy to my complex trait that would feed this one of the elementary, maybe low level traits that would feed into my complex traits. So IGF is not set up to do this project. It's very, very, variant centric, which is fantastic and great. But I think it's different from some of the functional activities we're discussing here. Right, no, I don't think anyone project's gonna do everything we're discussing here. I just wanted to raise that. And so then Nancy was asking about the multiomics project which we've just launched. So awards were just made this fall. So they're really just getting started. They haven't even had their first in-person meeting. And that's looking at really doing the multiomic measures in a set of specific disease cases. So with the disease individuals, they're small sample sizes. It's not the idea of we're gonna elucidate everything about diabetes with this one case. But coming in with all of the data generated centrally, a data coordinating center being set up and having all of the multiomic measures measured all together in the same individuals with a set of data standards that can help move, oh yes, longitudinally and in diverse populations. So there's a focus on diversity in the populations, longitudinal measures and having multiple omics measures at each of those longitudinal town points at those different individuals in a more standardized form that can help sort of drive some of the standards, drive some of the understanding, drive some of the resource and get at maybe some of these emergent project properties or ways to sort of say, okay, if you have those types of things that people always say are missing, the diversity, the longitudinal, the multiple measures in the same person, how does that move that forward? And in these exemplar case disease cases, what seems to be generalizable, a little bit in this point, and what really is specific? So what can, what do we get when we sort of look across and then also as we look more specifically, where do you have to have that disease specific or trade or context specific follow up versus this is something that can be done in this more generalized form? I guess one of the things I was saying is- Microphone, sorry. No, so people online, we have to microphone up for. I guess I was trying to make the point, not successfully, that we now have a model where much of the data is generated centrally and the analysis is diversified at the end. I think we should consider other models. You don't know that all the data done centrally is gonna serve the purpose, okay? So I think the reason why I said R01 purposefully is not just to enrich my lab, but just everybody's lab. That's the only mechanism by which you're gonna get everybody to think of very many ideas. I think what we have in front of us now is a complex set of problems. This is not the genome project. This is not the early variation projects or expression projects. And this, you have to leave it to both the imagination of the individual investigator, as well as their prowess in solving. They are more motivated to solve that problem than a data generation center. I know I'm gonna piss off some people by saying this and that's my purpose. No, and I'm gonna push you back. Not in the, I'm not a data diner since you didn't upset me, but I think the question I have about that, and I was talking about this a little bit with Nancy, is I do think that we wanna be encouraging the investigator-initiated work and we wanna be encouraging that activity, but then there still needs to be, we have the new data management and sharing plan. We're trying to get people to share the data, and this does still need to be a way that the information is being aggregated and shared. And so then in your model is there, how is that being handled, right? How is, like what are the approaches you take to that? I think I would push all investigators, including those outside those consortia to share data. That's a separate problem we have, okay? And NIH has all the machinery in its arsenal to push doing that. It's not only genomic data, it's antibodies, it's many, many other things. So I think that it should be a universal concern for all science. All I was trying to make is solving some of the problems in front of us as laid out. I'm not talking of only experimental paradigms. I think we'll require the imagination of each individual investigator to pursue it. Sharing of data, I think, should be uniform. And so just doing it in any major current model is not the only way to make the data transparent. You could do it. And say if you don't share the data within six months, we won't give you the next cycles funding. All right, somebody back there has been very patient. Go ahead. Okay, so I have one quick, maybe this is done already, but listening to all the talk between yesterday and today, there are two points that to me really stand out, which is the one is the biology and being in the GWAS world, but with the AP training, there is a massive amount of MPRI that are developed because they want to explain the biological mechanism behind the GWAS hit. But then when you listen to social people, they just say, what are you trying to explain? Some of these are just a certain bias or some of the signal are just picking up by affirmative factor that you're not going to explain, you're going to put the direct explanation on these with the biology. Now, okay, I think we have touched on this. There is an incentive of like, we have to survive in our life, we have to, they publish, we have to leave, we make it so that we chasing the number. That's another problem. And there are all these concerns that are chasing the number and the good work is kind of like disappearing. So I wonder whether there is possibility to incentivize work or actually have specific funding mechanism that incentivize combining those people like today where we have ecologists, we have social scientists, we have geneticists and basic science within a single ground to actually think about how we can tackle this problem and write a proposal that is picking up each of the point. Actually, if you say something that is stand out to me, which is the structure of the society with the racism that's been going on for decades, some of them are you picking up and some of them say, if you train a dog to be fearful that the generation will come with, like that's what would get spreading generation and then later on that's what you're picking up. Then I feel like maybe we are picking up some of these behavioral traits in our jurors and we try to put the biology into it, but it's just by having people from these different background and different world of thinking that can actually take us to like maybe scratching the surface. So I wonder if there's like mechanism that can actually specifically targeting these type of collaborative work. Can I just add, because I know that there have been some initiatives to do that in the past and I think one of the challenges of them that need to be more expressly thought about in these funding mechanisms is again, how do you not have it result in a situation where the LC person is just doing their LC critique on the side and not really able to move changes in terms of decision making that's happening on the part of the scientific researchers. So I think part of figuring out these funding mechanisms is going to involve finding ways and I mean I don't have the answers, I wish I did, but finding ways for those who are not a genome scientist to actually be able to make changes in terms of the decision making processes that's happening on the larger scientific team. Again, not to say that that that is solved. I think that saying we have an instance of that is not gonna say at all that that's solved, but I do think there is an effort and an increasingly to have the LC embedded at the planning stage, not at the after stage and really have that ability to change it. But let me just say I want us to get away from mechanism per se and to one of the questions. I just took federal assistance law, like it's very complicated why things are the way they are and a lot of them, some of them are changeable and some of them come from words in our constitution, believe it or not. If we're gonna take taxpayer money to produce something, think about what is the ideal structure of a project that produces that and what is it, that will help us understand complex traits in their context and in the environment and let's not worry about the exact mechanism or number of years, but talk about what is the collaboration that's needed, yes, what is the product, how do we structure things, not in terms of how they're funded, but how they actually come about and produce an end result. What is the, like 10 minutes left, what is the thing we have to leverage that we haven't answered the burning question, the how are we gonna get at making these connections across these levels? And while you're thinking about Jen's point on that, I do wanna address too, because I know you were talking, I wanna step away from the mechanism and this is a call to the whole room. I think we do have problems and we definitely have them at Genome and this is my acknowledging them, that we push for interdisciplinary research, but we don't come in and always recognize when you do that, we should all be looking at each other as equal partners and everybody is bringing something in. And so, as myself who was trained in epidemiology when I first came into more of the sort of genomic science part, there was a, oh, you're just an epidemiologist and I was like, I'm sorry, like I have a very specific training instead of information that you, that I need to bring. And I think that we all need to, like when we're doing interdisciplinary research, take that step back and recognize your do, we're not people aren't doing that just because it's like a buzzword, it's because everybody has something to bring and how do we set those partnerships up that that's being recognized and done. And I just wanted to acknowledge, because I felt like that was underlying part of your point in a way that I just wanted to acknowledge. I think the whole notion of the embedded ELSI model like is not in support of that in the sense that it's ELSI embedded in this larger picture, it's not in support of that equal power sharing. I just want to add one nuance in my question, which is I'm not talking about a model in which you have the ELSI person, but what I'm thinking about is actually project in which you'll have equal collaboration for a social scientist and then a geneticist and then an environmentalist, but really a project where in the proposal you see the contribution of each of those, each of the background. Thank you. So I had the microphone before you framed it as the big question, but I'll still offer in terms of support for the models. But beyond that, better sampling and practices around measuring context, it seems like that came up, but hasn't been mentioned here. And in terms of environmental covariates, how do we measure them? How do we, connecting to two of these concepts, the points of convergence? What are, if we get the UK Biobank and have access to it, thousands of variables are relevant to the environment, but how do we get to those points of convergence that are really essential? Something conceptually that came up at this meeting and how to, our phase or ideas would push us towards identifying those and both in the environment and in the endofenotypes. So Nancy mentioning inflammatory biology, cortisol levels, things that would be like key endofenotypes that are in that mediation of the environment and internal biology, I think would be a focus that, again, we heard, I feel it was coming up yesterday, but not in this discussion yet today. I was going to say something very similar to that, actually just a call for basic biology. And I think this is relevant to the comment about the Valley of Death and that correlational kind of statistical associations in humans might not get us there and in vitro studies might not get us there, but there's a lot of really cool diversity of biology and other types of organisms where there's actually opportunities to really understand how complex phenotypic variation arises over the course of development. So it's just a plug for basic biology in non-human animals. I have a question for Anastasia, I guess. So for a couple of years, there was the Genomic Innovators Award, the R35s. And I think one of the best things about that mechanism is it allows flexibility. And so you're not locked into something fan and given how fast the field moves, it's a little bit hard to sort of change course if you're locked into five-year R01 kind of mechanisms. And so I know it was discontinued and pushed over to maybe an R more of an R01 and I was wondering sort of what the thought process is behind that and if there's sort of an avenue for that to allow individual investigators to have that flexibility to adapt to things and answer questions that maybe are not within the frameworks of data generation or these large consortia. Yeah, so I'm gonna do a quick answer to your question because I wanna come back to Jen's point. So the decision to switch from the R35 to the R01 in that particular case had to do with the fact that although we had some really great people coming out of success successful from that, there were actually underlying issues with who was reviewing well and how that was going that we had issues with. There are questions about, and we're having ongoing internal discussions about, do we create an R35 in a different context in a different way because of the advantages that they have. And that is something we continue to discuss internally. And yes, there are advantages to that, but yes, that's my answer. And our last five minutes, we've heard about one paradigm that needs to be flipped on its head, right? Are there other paradigms that we need to flip in order to make progress in this area? And while you're thinking, I'm gonna remind, okay, go ahead. I think the nature of science that all of us have come through, we came through, all of us came through, doesn't matter what science, through a time of specialization, but we're not trying to tackle problems where it requires multiple skills. We can't be monolingual anymore. And you see this, you see people who were only doing say computational theoretical work and now doing some kind of laboratory work and obviously all biomedical scientists have figured out that computation is a part of their daily scientific life. So I think having some encouragement of all of your grantees that they need to, there's no reason why you're a geneticist. I think in solving some of the problems we've talked about yesterday and today, some of us have to become cell biologists as well, not in the way that we'll make fundamental contributions to cell biology period, but solving upper problems will require facilities in other fields as well. For other problems, it could be neuroscience. So I think that's an important step. I see that in my own work and I see this in the work of many others. So having some way of doing it rather than letting it develop just randomly or passively, I should say, may help the overall science that you already found. We're just getting the mic, go ahead. To piggybacking on Arvinda's comments, stimulate me thinking about how we've had some discussions with our graduate program of how do we teach enough social science to our geneticists that they'll be equipped to think about traits in this sort of era. And so that may look like curriculum development might work, involve funding partnerships with social scientists and geneticists to come up with new curricula to teach us just enough. I mean, we can't do double PhDs, but we could help get some of this fluency and the concepts and the tools of these allied fields that we don't have right now. All right, I'm gonna ask do any of our organizing committee have last ideas they'd like to put forward and also to thank our organizing committee? Do you all wanna stand up and we can give you a round of applause? Really hard things. We ask them don't talk about your latest research, talk about what we tell you to talk about so that everybody learns together on things that we need to be thinking about to make this meeting work. The presentations were superlative and the points that people brought up and got before the group were really important and I'm so grateful that everybody who attended attended and did such a good job at channeling what we hoped you would. And so thank you all very much. And so you know, the director of our institute has been sitting in the back listening to everything you have to say in this session and trying to absorb it all. And it's my great pleasure to introduce Eric Green. I wanna stand behind a podium, but I guess I have to. If there's a handheld, I don't know, I just, I don't feel like I warrant a podium. Is it working? I don't have particularly formal remarks, closing remarks to make about a workshop I didn't participate in. I would have liked to have been spend more time here than the last 20 minutes. I do have a few remarks I do need to make. I figure there is time. It's not supposed to end up to 245, right? So I can riff. I always feel like if people come in and you know, get on airplanes or trains or wherever I drive here and the least we could do is to give you some opportunities to ask questions. I'm happy to answer questions, but also to just give you the latest there a lot is happening. The reason I missed 99% of this workshop is I have been sucked into a vortex. Some of it of forward movement and not so forward movement on a lot of things that actually will affect all of you. Some negative, some positive. So I guess I come here with good news and bad news. Do you want, I'm late on a Friday. Do you want the bad news first? Do you want the good news first? Want the bad news first? Put on your seatbelts. The budget is really gonna suck. I mean, I don't have any way to sugarcoat it. I just think I'm sure you're interested. I'm looking at major leaders of major institutions here and I know you know this, but there is nothing on the horizon that looks good right now. And of course, why is this week any different than any other week? Well, we at least now know that we're on another continuing resolution now with part of the government in one deadline, part of the government on another deadline and for NIH, it'll be February 2nd. So now we don't know our fate until fact this actually is also mostly bad news because the further it gets into the fiscal year without us even knowing if we're gonna have a budget or a year-long continuing resolution, I cannot tell you how crippled we become programmatically in making decisions because we don't know what we're making decisions about and with less of the fiscal year to get money out the door. So it actually, you know, it obviously sounds like it's going from bad to worse and it could even get far worse because of gosh forbid and for February, if we would take a budget cut, it would be devastating. But we are at best looking at a circumstance, we believe and I'll get to our new NIH director in a minute and she does have a lot of conversations that give us a lot of insights and sharing with us now of why it's not looking particularly good is that we probably are gonna be flat but flat is not flat, it's not even close to flat and it's also looking like it's not just this year, it's probably looking at a couple of years to get to the other side of election, all that. So sorry to tell you this, what it does and part of what the vortex I get sucked into is there's a tremendous amount of hurry up and wait going on and of course we were waiting, waiting, waiting and now we actually know that we're in a CR until February we get, I mean, poor Carolyn had to be on a budget meeting with me today for an hour. That was only one of four hours of budget meetings I was in today practically. It just allows you to plan a bad year as opposed to sit on the sideline waiting to plan. So I wish I could give you good news. With that said, the silver lining is please don't stop and I know you weren't. Please don't stop being strategic. Please don't stop the science. If we have to take some great ideas and just put them on ice for a year, we'll get there. We don't want you to stop being innovative. We don't want you to stop telling us the greatest things that we could be doing. And I'm as impatient as any of you but we're just gonna have to be patient to some extent and wait for the budgets to improve and most importantly vote and get all your friends to vote because we need more people who are enthusiastic about science at a time where we've lost some of our strongest science advocates in Congress. They've retired and now we need new ones. And actually also talk to your local folks, be local because tell your stories, please go back. I know some of you do, I'm looking at faces so I know we're quite involved. Get members of Congress to come and see the great science going on at your institutions and tell them how incredibly important NIH is. We need to rebuild the advocacy that we had that got to the point where they retired. A lot of them were stepped away from Congress, a lot of them. So that's the bad news. The good news is and it is interrelated is that we have an NIH director and which means, and by the way, I would say that Larry Tateback for the last two years has been phenomenal as an acting NIH director. So it's not like the place went bad. It's just, and I think all of you know because all of you have been in an institution in case you have people in acting positions and they could do a lot of things and then there's just a lot of things that just get on hold. And Monica Bernalie is now in place. She got sworn in last Thursday late afternoon and so she's eight days into the job and it's like the floodgates are opening. All the things that were sort of a little bit on hold, boom, boom, boom, boom, boom, boom, boom. There's a lot of things happening. And including one today, which actually caught me by surprise, maybe Nancy knows about this. It's probably a huge loss for Nancy. Sucking up to a new chair. Yeah, yeah, you have to a new chair. Was she actually your chair? Yeah. Oh wow, so for those who don't know, President Biden today, or I think it was today, nominated, it's Kim, right? Kim and Raffaul. Yeah, Raffaul, so there you go. And as the new, or has been nominated to be the NCI director. And I guess that was Nancy's chair. I had to figure it was a big colleague, I didn't realize. And I hear already from people like Josh Denny that she's phenomenal. And I don't wanna put you on the spot with the cameras rolling because of course you're gonna say good things about your chair. Great thing, really great thing for NCI and for my age. So there you go. So there's good stuff there. Sex for me. Yeah, sex for you. So, you know, NCI is a very important, the other thing I got excited about that is she blows genomics, right? That's what I, that's great. Green, there you go. Okay, good. Now we're back. So in any case, and that is, without getting into many details, some are pedantic and some are really exciting. It's just, there's just a lot of, a lot of energy all of a sudden because Monica, as you expect, has been on deck for a while, finally got confirmed and she has an agenda and she's not wasting a moment to get people mobilized to move things forward. So lots of things that have sort of been a little bit on hold, a lot of energy. We don't even have that many vacancies but we have a few vacancies. She's ready to move forward to do that and some initiatives. And also to have someone who's really at the helm to help us navigate tough waters of budget. And so there's just a lot of things and many of you who serve in advisory roles, I'm sure you'll, over the next year you'll hear more and more about that. Permanent leadership for the NIH. Okay, so that's the big, I mean I will close out with a few general comments. Any questions? Because again, I don't have to use up the 15 minutes but I'm always happy to answer questions in situations like this. Some of you I see more frequently or some of you, a lot of you I saw at ASHG fairly recently. Any questions? Okay, so I mean what I would say, I wanna thank all of you for participating in this meeting. It has been an area that we knew we wanted to unpack with brilliant people like all of you and folks who have been joining us virtually as well. We really saw the meeting as valuable to explore how best to implement ideas that were not surprisingly laid out in our strategic vision that we published in 2020. Specifically, we talked about the importance of having an ambitious and compelling research project of quote determining the genetic architecture of most human diseases and traits. Of course, something of deep passion to all of you. And we want, and we continue, I'm sure this came up and again I always feel strange making remarks at a meeting I didn't participate in. And I know there's other institutes here or have been tuning in virtually. It's tremendous interest across the NIH and they are absolutely looking to us and others to help guide how are we going to really tackle the complexity of most genetic traits that are relevant in human health. So it was really an area that we wanted to make sure that we were, even though we've had workshops in the past and we've touched on it in various capacities, we wanted to make sure to have a dedicated workshop to stimulate the discussion and engage most human from what I heard at the comments I was at least able to listen in on it and it sounded like it's been a productive couple of days. I did hear a little bit from the discussion that Jen just left that gives us lots of ideas to think about and lots of ways also to not only programmatically tailor some of the things to sort of maximize our return on investment in this area but also how to intersect with other NIH institutes and centers. And again, as I want to point this out, maybe it's obvious, it's an opportune time for you to be helping us think about how to interact and think about some of the other things going on at other institutes. Let me expand that point because I think it's a really good one. Looking around, most of you we know are funded or very typically funded by multiple institutes. Maybe you're funded by us, maybe you're not but you're also funded by other institutes. You actually get some insights into some institutes from a different perspective than what we get. And what I would say is, and of course, you see the differences and the similarities and you've come to learn how we're really cool and most of them are not as cool. No, I'm just kidding. But what is really important is there are incredibly fresh eyes at the leadership table at NIH and it's not that there's anything wrong with the non-fresh eyes. I mean, I've old eyes, right? I've been around a long time. Francis Collins was around a long time. But it's amazing how many new institute directors there are. I would point out, I think there were four that came in during the pandemic and three more that have started since the pandemic. Again, it's like seven new, and now you have a new NIH director and one of them is the now NIH director. And lots of the usual ways of doing business and lots of the ways are just being questioned. Not in a negative way, but just in a way like, well, why have you always done it that way? Have you ever thought of doing it this way? So there's just a lot of new looks where questions are being asked. I can tell already, I'm only eight days into this with the new NIH director, I can just tell the questions that are being asked are really helpful for thinking about, is this the best way to organize it? Is this the best way to do this in today's life? Maybe it was perfect 10 years ago, but are there ways we should be doing it? By the way, in the data domain, expect a lot of really deep critical thinking about how NIH does business in data with a recognition that it needs to improve substantially and I mean that in a very broad sense. So the thing I like about workshops like this, and I'm sure I'll be hearing about it from my colleagues who will give me summaries and et cetera, we'll be discussing it, is I hope occasionally things came up and I heard a couple of things come up even the discussion, is it doesn't have to be business as usual. Don't be shy about prodding us because I'm gonna be prodded by others. I know at NIH about why are you doing it this way or maybe we should try to organize things a little bit differently otherwise. And you also have to do that when it's fiscally tough because then all of a sudden you gotta think about, well, maybe you gotta think about other ways to do it when money is really tight. So continue to do that, we hope we got that and I think we did hear some of that. Last thing is just to thank, we already heard, thank the organizers and the organizing committee and I know Nancy, Lauren, Molly, Peter in particular. Our folks from NIH, Xander and Jake and Ismail, thanks to all of you, thanks to our AV crew at our web team for broadcasting this out. People have been listening to it. We did video capture it and so forth. And I think that's it. Carolyn, Zander, anything else? All good? Okay. Oh, Sarah, okay. She won, nonetheless, but Sarah, thank you. Anything else from anyone? If not, it's beautiful out there. Safe travels, I know some of you are heading to airports and so forth and you will be in touch and thanks so much for giving your time to us. All right, thank you very much. Thank you.