 Great. Thank you, Howard. So for session one, which will last an hour, we have two keynote addresses. One from Nancy Cox and one from Mike Snyder. Each speaker will have 20 minutes for their presentation followed immediately by 10 minutes of Q&A. So our first keynote speaker is Nancy Cox. Dr. Nancy Cox is the director of the genetics within the department of medicine at Vanderbilt University Medical Center. Dr. Cox has a long-standing research interest in identifying and characterizing the genetic component to common human diseases. Her current research is focused on large-scale integration of genomic with other omics data, as well as biobank and electronic medical records. Nancy's talk is entitled multi-omic integration to understand health and disease. Nancy, the floor is yours. All right. Thank you. It's a real pleasure to be here and I'm very excited about the opportunity to share some of our thoughts on where we might move forward more rapidly in these spaces. I think this moving into really health and disease, moving omics into health and disease is where we need to be. And it's an astonishing thing to look back at the last strategic plan and how much black NHGRI got for focusing so much on the genomics in medicine. We weren't ready for that. We weren't ready for that. And then how much further we actually got in that whole space by the end of the strategic plan. Genetics is part of medicine in a much bigger way now. And it is coming faster and faster into medicines and medical centers and hospitals. And that creates some truly unique opportunities for us now, because as more and more clinical genetic data become available, every medical center, every hospital becomes the equivalent, as I'm going to say, of a high throughput phenome screen in the context of the omics data that become available in those centers. Full advantage of what we know is coming. We need to be proactive in creating the foundation for using all of this genomics data that is being generated in the context of health care systems, which will have electronic medical records at their levels, and of course more and more of their patients with wearable devices collecting all kinds of additional information. We as a community need to pay more attention to the education of medical center leadership in not just the risks of of conducting joint research is kind of more shared kinds of models of research. But in the unique opportunities that gives medical centers to enhance system. So, we need to educate do more education of medical center leadership on the opportunities, they'll leave on the table. We need to perceive the risks as being higher than the benefits because the benefits stand to be enormous, if we can get a handle on this. And of course hand in hand that's a lot more education of patients and physicians. It's really important to remember that just barely half of practicing physicians in the United States have ever had a course in genetics. So, a lot of what we're going to do in genomics coming into medicine. It's going to have to have a lot of implementation support, you know, kind of automated consults for for even ordering tests, and certainly for interpreting results, a lot of baseline support for physicians. We need to be thinking now about models of governance that that can allow these, not just, you know, clinical data warehouses with the diagnostic information, test results, medication histories, all of that information. But also the omics sandboxes that will be created as genomics and other omics technologies do come into medicine models of governance that that allow us to look across systems to pull more information together, which goes hand in hand with enhanced kinds of security, not just for patients, but for the hospitals, this, the hospitals are just under siege now. It's an incredible thing. And COVID is like enhanced it in odd ways anytime any of our medical center people are on a network show talking about COVID and our response to it attacks against our hospital system go up 20,000 fold over the next several hours. We need, if possible, to be able to truly federate queries across these systems. And so the education, the governance and security goes hand in hand with what we, we hope to achieve in in being able to really put this information together as fast as possible, so that we really enhanced learning how to use genetics, learning how to teach the use of genetic genomic and other omic information in the context of healthcare systems. And if you know if we just think about what, what kinds of multi omics we need to, to prioritize integrating. It's a crazy thing but but we still have to do to truly implement knowledge we already have genome by genome which is a legacy of this crazy dichotomy of common variant studies in more common diseases and rare variant studies in more rare diseases, because we just, we have to do just genome interrogation and full scale genome interrogation with full scale genome analysis. Whether we're talking about common diseases or rare, we all know breast cancer is a great example of a common disease that where common variants are contributory, where there's a polygenic risk that has meaning, and where there are rare variants that have relatively large effects and, and integrating that information is critical. To some extent there's better and better analysis of whole you know, in the common disease space, where, where, where people have really been trying to working to pull this together. Honestly, the end of one space is lagging, and a part, I'm sure, of the failure, the success rates being still mostly less than 50%, is a consequence of failing to recognize the common variant contribution to people that have what could also sometimes be a consequence of rare infection. I think we can do a lot better job really quickly in that space with, with modest investments up front now. And of course, you know that all of the, the rest of the omics investigations it was great to see how much investment there really has been in that space but I think, I think, by drawing a connection to a backbone of genetics, we have some additional opportunities. Think of hospitals, all medical centers as being about to be biobanks of the richest sort with years of electronic health records data on people, as well as genome variation for the possibility to impute in transcriptomes, metabolomes. But in the not distant future, and I think because of the opportunities for transcriptomes, metabolomes, proteomes, epigenomes as being really important biomarkers, we will have more and more of that directly measured and, and be able to do much richer high throughput human phenome screens with these measurements, more and more things like not just all of the medication history test results, vital signs, but also imaging data procedures that people have just hospitalization information, but the wearable devices, and the data that's now being collected from patients on a regular basis, sometimes daily, sometimes weekly on progression of their Parkinson's. More and more in the space of neuropsychiatric disease, wellness states, especially for those diseases that wax and wane that have real cycles. The opportunity to, so I'm already blown away at the early data on the utility of what's collected in wearables, and and, and these kinds of regular phone information from patients integrating that with the electronic health records phenome is an astonishing and rich source of increased understanding about medicine. That geneticists really need to have so I, the phenomic, the, the depth of phenomic information we have really offers us a lot of opportunities, unique opportunities now. The other thing we haven't talked about as much but if we have genome information on people in whom we also have measured transcriptomes metabolomes epigenomes, the opportunity to take advantage of the delta. People physicians especially think of biomarkers as dynamic measures of disease progression, but biomarkers are effectively as heritable, sometimes more heritable than transcriptomes metabolomes, and these other measures that we know are easily predicted from the genetics. The opportunity to simultaneously consider the genetic and static component to these features, and then use the genetics to highlight the dynamic ways that existing biomarkers, but also the omics biomarkers. Tell us about the body's response to this homeostasis and the descent into disease is something that we should grab a hold of with both hands. And I'm going to come back to that because this disconnect of the medical establishment in understanding how heritable. Their biomarkers are, I think it is an opportunity for really improving medicine in the short term, as we get more and more genetics into healthcare systems, because we can begin to separate the, the sort of innate component of this from the, the truly dynamic aspects of our body's attempt to return to health and homeostasis that I see this as a huge new opportunity as more and more omics come in, but one that we can start with using existing biomarkers. And, of course, genetics will come in slowly just, just as for many of us with biobanks, only a part of our data set actually has genetic interrogation, but we have millions more with the EHR phenomic data. And there are so many untapped opportunities for enriching what we learn in the genetic space by taking hypotheses forward in the millions more people, the tens, hundreds of millions more people that we have these very rich EHR data on, and especially in the drug repurposing space. This is an untapped resource that if we were able to get better federated queries across existing electronic health records, thinking about repurposing, drug repurposing could really be taken to a new level. I think there are just tremendous opportunities in, in taking what we already know from the genetic space, but then driving that into data where we have information on whether drugs that hit this piece of biology that we think it's related to Alzheimer's delays the age to onset for Alzheimer's. This is, is, if we could just get this together, these are the kinds of things that could be done today with information coming out of genetic studies. The, the use of the phenomics information to get to new kinds of phenotypes and allow us to do new kinds of things. I think it's tremendous and, and some recent research from, from Doug Rudifer and, and folks at Vanderbilt on how to use the phenome signatures in the clinical data to really identify patients that have all the characteristics of people that have been sent for genetic testing, but who themselves have never been sent for genetic testing could be a game changer in getting better information to insurance companies that would allow them to be much more proactive in thinking about reimbursing. If they understood completely the sensitivity and specificity, because we now would have much more system systematized way, ways of bringing genetics in. I think, again, looking at the bigger picture, it's likely to be a faster way to, to move genetics into medicine in a way that we, we all think it needs to be. And the, the machine learning approaches, even simple things like the phenome risk score. There are clearly undiagnosed pace patients for all kinds of Mendelian and syndromic conditions that that we can see through even things like simple phenome risk scores, and another part of this space is the laboratory values, which are complicated to use at scale. But, but recently, a group from Vanderbilt published a whole pipeline, enabling the cleaning of medical center based laboratory values that can get us get you really to the quality of data the same quality of data that we're getting with research performance you can see that in the context of the heritability and genetic prediction analysis. So, so all of a sudden we have literally hundreds of laboratory values that we can use to enrich this, this space. And if we think about how do we prioritize really bringing genetics and omics into medicine. I think the first do no harm mantra is a good way to think about it, but that means also potentially undoing harm that we've wrought. And, and a part of that really is these laboratory values that are to physicians shockingly heritable. It's not shocking to geneticists who know how heritable most things are. Most labs have significant heritability, and of the labs that have significant heritability, most then have population differences in population mean differences by self reported race ethnicity categories, and this has generated an underbelly of really problematic. Essentially institutionalized racism in the context of it being much more common for popular for individuals from us minority populations to be billed multiple times for repeat testing to be under diagnosed. To have unnecessary procedures to miss the opportunity for early diagnosis and treatment of important diseases, and I think that there really is a huge opportunity for for geneticists to help in in trying to undo some of the damage that has been done in this space, because physicians truly did not appreciate how heritable these laboratory values are, and the fact that they are heritable means that you know local adaptations and different continental arenas can lead to average differences in these values, the increased variation in Africa, not uncommonly leads African Americans to be outliers at both ends of distributions more frequently to be to be billed for repeat testing on both ends of some of the laboratory distributions. But I think this is an area that we, we haven't appreciated as contributing to I think some of the health disparities we observe that have been historically characterized as minority populations not having as much access to health care. I'm afraid that are not insubstantial component of that is that they are absolutely taking good care of their health, but because of of having higher average values for means that that the reference ranges that are being used are are not catching them early enough in their disease states. One minute Nancy. Yep. Okay, we're, we're, we're good. So, as you all know there are no simple, simple solutions to complex problems but we really shouldn't go on this way. Once we become aware of the magnitude of these kinds of problems using the genetics and other omics is a great way to get an environmental factors that we don't know about know how to measure can't get at. We can use genome variation and the other omics to better see this whole picture and the return on our investment, as we discussed new biomarkers, but genetics as biomarkers and the delta on existing biomarkers and omics biomarkers, using all of this to understand new types among diseases, signs and symptoms, as well as non genetic factors, making Pliotropy our friend, as well as our enemy, it's, it be devils us in Mendelian randomization, but it can be a huge help in a good way for thinking about better informed clinical trials with earlier onset Pliotropic phenotype. When we're thinking about something light onset like Alzheimer's. When we reoriented the trees for diagnostic trees to reflect genetics, it would be much simpler to be working for new uses for existing drugs that the repurposing angle, and, and new drugs would would would be more easily pursued because it would be more obvious how more people can benefit from the drugs and really getting to disease prevention with this with interrupting. It's not just that genetics is and should come into medicine, it should be the backbone of medicine, it should be an organizing principle that makes everything better, simpler and faster to do right and as has been said to improve health. To keep people healthy to keep them to bring them back from that slide into to disease states through dyshomia stasis. So, with that, I'll close and happy to take questions and discuss. I'm going to stop sharing the screen. Yeah, people could show their videos during the discussion period that would be wonderful. So the first question comes from Terry manilio per your point about diseases that wax away MS lupus inflammatory bowel disease. Is that a particularly rich opportunity for identifying short term predictors of exacerbations and potentially extrapolating that to predictors of onset or progression. I think, I think there are incredible opportunities in that space and we really think we have a lot of the basic research that we need to in order to do this, and it's just getting into the right clinical spaces to get a better understanding of the full opportunities. Great. Next question David Craig you want to unmute yourself. So, you know a lot of experience with the learning how the wearables and that this is their approach in this and the idea is, we need to get the genomics or into the EHR, and then we need to get federation which means we need to open everything up. You also mentioned, we get these attacks in our system. And my question is, are the phenomics communities, people who do wireless, are they taking a different approach because I hear them talk about wanting to yank everything out of the EHR in work in a platform where they're basically free and clear. And then I'm struck about how fast they are making movements and I'm wondering whether my wearable will be more informative than my genomics multi marker in 10 years like will we get crushed by playing by a different set of rules. I think the, the rules are so rapidly evolving in this space that I just trying to keep up as a challenge. I suspect there's actually going to be more more rather than fewer rules in the not too distant future that relate to wearables. I've already seen a pullback in the context of projects that that I know about at Vanderbilt that use wearable devices. It gets a little bit more complicated. Now, over time, because of the, the complications of providers that used to have much more data at their disposal, people having opted out of some of those ways of collecting the data may need to opt into more specific uses that that have value in the context of health care and medicine. So, it's working both ways I think with the wearable devices on on how that how easy or hard is going to be to collect large scale information there, versus the health care systems and people potentially being able to slide slice their whole EHR and be able to have it used in the context of shared health initiatives. I think, but, but the sooner we, we really get more education governance, and, and people appreciate the opportunities that this space really provides for learning health care systems. I think the better will off will be as the clinical genetic data come in for for really accelerating the learning curve. Next, Greg Gibson, Georgia Tech. Hi Nancy, fabulous great to see you. Thank you for that. I was wondering if you could expand on your comments on the n equals one genetics and what you see the role of multiomics being so I imagine you're thinking that transcriptomes and methylomes get you close to the phenotype. But there must be enormous challenges in clinical implementation and proof of pathogenicity. And then that relates to the education of the community as well what are your thoughts on on on educating physicians. So, I think that, you know, my husband is a physician, and he says he's too old to learn how to order a genetic test. He wants to do a consult, and he knows how to do a consult now to order genetic tests that he needs in his endocrine practice. And he expects that the return of the information will be will make it very straightforward for him to understand what he needs to do for this patient. So that I think is what physicians want to know is what do I need to do to take the best care of my patient. And, and that is how we have to provide the, the return of information is the emphasis on what this means for your patient in this context. And I also think the, the opportunities in the end of one space for enabling medical geneticists to, to better tap existing knowledge, even with just, I mean I got to say so many of these an hour whole genome sequences. You know, one of the things that we're looking at is you can predict out the transcriptomes for each person from a whole genome. And we already knew from Vanderbilt's Biobank, the people who are outliers, greater than two standard deviations above or below the mean three standard deviation we can set you thresholds for every one. People who are outliers with the number of genes where they're predicted genetic expression is, is, is far off the average. They have they accumulate way more clinical diagnosis and, and in, and many of them resemble and have one patients. And we're looking at, at undiagnosed disease cases, just using that that information from our Biobank, which is summary stat data that can be made available for queries. So that, so that physicians looking at a whole genome can get a handle right away so is, if their patient is not an outlier for, for genetically predicted the expression of a lot of genes. Well, yeah, I mean I, you know, there's maybe more likely to be a rare variant contributing, but when they are. They're not different from a lot of other people in medical centers that have a lot of accumulated a lot of health problems. And that would be a good thing to know, because we can often show that the outlier predicted expression of those genes is associated to precisely some of the phenotypes that are seen, you know, at least statistically. So I think those kinds of that's not even an additional omics to what's being obtained, although I love to see to look at the delta on genetically predicted transcriptomes to the measured for opportunities to hone in on the the pathways that are most affected by potentially whatever rare variant is a driver of those phenotypes. So that, so I think there's, there's some terrific opportunities in that space that would be straightforward to make widely available to the medical geneticists who have to work in those end of one spaces. Thank you Nancy, you have multiple questions in the chat and raised hands, maybe you can address offline but I will, I will, I will do my best to address those as we proceed. Okay, thank you so much that's a great kickoff. Our next speaker our next keynote speaker is from Dr. Michael Snyder, who serves as chair of the Department of genetics and as the director of the Center for genomics and personalized medicine at Stanford University. His laboratory has developed many technologies, which have been used for characterizing genomes proteomes and regulatory networks. He performed the first longitudinal detailed integrative personal omics profile I pop of a person and his laboratory pioneered the use of wearable technologies smartwatches and continuous glucose monitoring for precision health. Dr. Snyder will talk about personal profiling using multi omics, like All right, well great thanks for including me in this meeting it looks awesome. And so I want to tell you it's going to pick up very nicely I think for what Nancy was presenting about using multi omics to be able to profile people's health and catch disease early. And it'll show some things you can do with it that go way beyond what you can do for the genome. People tend to study one thing at a time say the genome or the transcript them really what they're doing is we're getting just a few pieces of the puzzle. And I think we would all agree that it's very useful to get the whole picture as best you can. And to me that's what multi omics does it lets you capture lots of information so you can see all kinds of detail you might not otherwise see. And I can tell you we do a lot of studies, and if you may know and the one I'll talk about is profiling people's health with multi omics but we do work in the early cancer space around exercise all using multi omics profiling for trying to see changes and I guarantee every single contributes things you would not get from single arms alone and of course when you're looking at things like exercise what have you. Obviously that goes way beyond the genome the genome is fantastic for multi omics profiling based with the additional information again, but you see much better what the whole picture is. So diving into personal health which is one of my favorite topics. You probably all know your health obviously it's impacted by your genome for rare and complex disease, but then all these other things contribute you know your various exposures pathogens by mental stress food exercise. We can tell these things some well like activity or about the wearables. Others are still pretty clunky food tends to be missing a certain amount of data stress believe or not we're getting better at so we can quantify this but we can even better than that quantify its by doing deep profiling on people. And so they're now advanced technologies obviously the sequencing technology that we're all familiar with but mass spectrometry is going way beyond what it used to be 15 years ago so it can follow. You know, thousands of analytes and people's blood and you're in wearables came up before and I'll talk about all these things. I'm going to let you get a much clearer picture of people's health if you will, that was ever possible before. And so many of you may know a number of years ago actually it's about 11 and a half now. We set up this personal omics profiling that was mentioned and started out on me but then we assembled a cohort around 100 people a little over eight years ago that we've been profiling and and in its most deluxe form will measure 14 different homes will measure the genome once and then we'll measure the these other various homes this is tends to be DNA methylation. And they're typically measured out of blood components because that's pretty accessible like your immune cells peripheral blood monocyte cells we tend to measure epigenome transcription and so on and then other things will measure out of blood and urine for example metabolomics and so again for in its most deluxe form will measure all these different things will also, but for most people it's typically about eight different alms. And then on top of that we have clinical tests and questionnaires, we have some advanced tests that tell us a little bit better what's going on for certain individuals. And then the wearables we added on about seven and a half years ago or so before Apple Watch I'll talk about that just briefly at the end. So one aspect of all this is very very deep, multiomics profiling. And the other aspect again builds nicely from Nancy's comments is the longitudinal nature we're following people over time. And then even though it's a small cohort the goal was to see you know, what does it mean to be healthy what can these profiles tells about health how does it differ between different people how does it change over time in an individual. And perhaps most importantly related topic of the session, you know how do, can these advanced technologies that are being used to make these measurements can they be used to actually better manage people's health, then go beyond what we're doing so I forgot to say but we're profiling people three every three months while they're healthy and then if an adverse event comes along like viral infection, we typically take five to seven additional samplings. Okay, and so again we've been doing this for quite some time. And so just to cut to the chase on the last point about health. It turns out from these 109 people we had 49 major health discovery 67, hypertension. And so what's powerful about this is several things one is that no one technology found the covers, well, covers a wide range of space so human ecology cardiovascular disease metabolic, other areas, and no one technology actually found this this is the power of genomics I'll go through this a little more in a minute, but sometimes for example, it would be genomics other times be imaging usually it was combinations of data was very, very powerful. And just give a few examples, you know we did catch some with early lymphoma to people with these pre cancers are called they can convert to aggressive cancers and so finding them early. So to be a resource heart issues one from genomics one from wearables, and so on and so forth so we actually found a lot now. This is probably a higher number than you might normally expect because the average age of the person entering our study was 53.4 but it was a pretty wide range as well. So, so nearly half folks, but the other important point out of all this is that every one of these things was found pre symptomatically that is to say we made these discoveries before people had symptoms which is a big deal, because then you can take preventive action which is really what's lacking in the healthcare system today. So, just to give you a few examples genomics certainly obviously was really, really important to help people have these Mendelian kinds of problems the ones of red or the ACMG genes for those of your follow this these are two means or two individuals with these and others were patients and other things and three of these people had actually health issues associated with it so one person a pretty young guy had a mutation in this gene and which is cardiomyopathy gene and turns out how to heart defect did a follow up. And sure enough actually has heart defect as all drugs now this individual also fairly young did a whole body MRI follow up and once you know I had early thyroid cancer that part was removed was able to keep most of thyroid does not need thyroid hormone therapy. So, and then there's nine individuals coming in study that were were thought to be type two diabetic it turns out one of them is a modi person and was on the sub optimal medication because they were modi which is treated a little differently from a typical type two diabetic. Here's some other cases we found one person with the early lymphoma was picked up with imaging but also had some elevated blood markers from the proteome and the combination was quite informative actually and that's the other thing about this way do multiomics. It's getting multiple measurements that are all telling you something's off because I think people often see that you may have experienced this yourself. You can be in the normal range for clinical markers and following along pretty nicely, but suddenly you'll have a big jump up in a marker and still be in the normal range but it's not normal for you and that's a big theme out of all this. Here's an individual with this pre cancer also pretty young person had two outlying measurements had to two samples at the time there are 1000 samples here everybody else's sample measurements were here for GM. This persons were here this pick up with proteomics. And again was elevated follow up said yep too many IGM producing cells and now is is is being followed pretty carefully. People have plaques in their arteries this company imaging, and this is actionable information these folks I'm one of these this is not me but mine looks a lot like this and so recommendation is increased statins and that was done for me and my practice shrinking so that's a good thing. This is the kind of thing you can learn. We're also setting up predictive scores for using genomics plus other markers that actually are we are better than the framing him, where the more recent scores for risk for cardiovascular disease. From what we found I won't get into that we're doing quite a bit of that. But many of you know a pet interest in type two diabetes because I myself I'm type two diabetic. So just coming into the study nine individuals were thought to be type two diabetic to others were diabetic who didn't know it. And then plenty of people were pre diabetic turns out nine out of 10 pre diabetics. Don't know it and I guarantee a lot of you listening to me right now pre diabetic and don't know it. And this is a big deal because 70% of those will become diabetic in their lifetime so again we think getting early information before symptoms is a big deal. We are very curious how people become type two diabetic when it was one of the main reasons we did the study so nine people became clinically diagnosis type two diabetic. And others became pre diabetic some others had diabetic range values but weren't officially classified. But so we're kind of curious how do people become diabetic do they spike the way there as though something triggered it or do they just gradually get there what what's happening and so for the nine who became clinically diagnosis diabetic. Two of them got there in a classical way they gained weight and their microbiome diversity dropped and they, you know their glucose went up and they went on that form. And then the other seven though didn't gain weight. Five of the seven just gradually became diabetic either through fasting glucose measurements or hemoglobin a1c this is in graze hemoglobin a1c six fives called diabetic for hemoglobin a1c and or fasting glucose depends on your reference I think 126 is the number these days. Anyway, some people gradually got there that way. Others through some a test called oral glucose tolerances. Interestingly, two people got spiked their way to diabetes this one's me I'm one of the two. My genome predicted it from a complex disease thanks to toll butte, and then I was following my glucose because of that pretty carefully and ironically when I went in for an insulin resistance test is me I was very elevated. And then later got measured sure enough as officially classified as diabetic I know this looks like a little blip. That's actually about 10 months is eight years of data for me. Anyway, in my case it seems have been triggered by a viral infection of pop. Because when out of control after a nasty respiratory sensation virus very relevant to the pandemic. For those who follow the stuff covered. There's an increase in symptoms type two diabetes after covered infection by something like 4%. It's a fair number so this is the first demonstration of association of viral infection with type two diabetes. I actually got under control by lifestyle change with food and exercise. It was boringly low and actually I stopped looking at ironically someone looked at here. And then I was reclass someone looked at it and said well Mike yours 7.0 and yours at the time hemoglobin anyone see, which is diabetic 65 is the diabetic value. So I guess what I stopped running here but I also had a second viral infection so it's not 100% clear what was the viral infection of running but I did start running I got it down never down to this baseline. And it gradually went up and and I won't get into all this but he switched from running because it wasn't. Gradually still creeping up and then switch the weightlifting and that failed so anyway, then I went on that form and like most people said you should have that beginning guess what I'm a non responder. So in the end I went up all the way to 75 actually and then about a year ago when I switch I got more data basically and I realized that actually make insulins just fine. And I'm insulins sensitive so my cells respond but what I don't do is release it from the pancreas which you can figure out by this glucose disposition measurement. And so it turns out there's a drug for that it's called repigulomide, and it actually works pretty well on me so so this is a classic case of more data gives you precision diabetes you can treat it with the right drug on the right person and now I use the two and neighbors are also pretty good. So bottom line is the multiomics profiling and you could say well I should just include cost that's nice to know but we actually now have measurements believe it or not the best measurement for insulin resistance is a six hour it's called SSPG assay we now have a multi analyte assay that comes from all the emics profiling 12 analytes is pretty close to as good as a several thousand dollar SSPG test so so we would argue it's great for discovery these multiomics and it's a combination of microbes metabolites and a few proteins. Alright, some of the other things we've learned from this I'll show you some of the fun stuff we learned is really relate to what Nancy was saying and some of the questions game up during her talk which is no matter which only use we can follow people over time it turns out. These are fasting glucose measurements your arms are very personal, and they're actually pretty stable so this is 12 people with at least 10 healthy baseline measurements. I'm the light blue guy six years of data for me you can see my side of kinds all cluster together each thoughts of different visit, each colors a different person. Brown someone else read someone else so everybody's measurements they're very stable and they're very personal, even transcriptome if I showed you this and 3D the third principle component you'd see they're separated. And so, and the other thing that's pretty powerful about this this is a big deal. If you want to go viral infection or other we have done weight gain weight loss and most recently we published an exercise study. You can shift your omics profile half your molecules change when you run to your VO to acts, the VO to max, but you will still cluster your samples will cluster relative to you, not relative to whether you ran or got a viral infection so the point is that for the there's some exceptions but the point is that these profiles are stable. And they're actually even in the perturbation, you look more like you than anyone else. So why is this a big deal well, it means that when you're trying to do studies of disease versus normal. If you just grab cases and control you need thousands of people to see that signal of disease versus normal. You're looking at the same person it's very very easy to spot because you see the shift from the baseline so it's easy to see your own shift and this is why personal omics profiling is so powerful that you can't miss a shift from your personal baseline into an early disease state, but you will miss it if you try to compare your disease state relative to 500 healthy people. And this is a hard concept. I think you lost folks have a really hard time with those based on reviews we get all you need 1000 people you don't when you're doing personal omics profile. You just need one you. Now you'll learn more generalization if you get more folks. I'll show you a few fun things that you can learn from money on this profile that you will never get from your just your genome. One is here's a simple thing we asked well, you know how many seasons are there out there the biological seeds how many, you know for your molecules how many how often do they, you know, change you know arbitrarily we've been taught since we were little that four seasons winter spring summer fall. And if you think about it well who says there's four and who says there's three months each it doesn't make any sense, maybe they're 15 season maybe there's three I don't know. We can let the data tell us so by profiling these folks we again had 1000 measurements these are the folks and turns out they're fairly evenly spread over the year. And then we can say well what kinds of molecules are showing seasonal patterns out of the, I should say microbes and molecules we're doing microbiome here gut and nasal. And it turns out there's about 1100 even though most molecules are stable there are 1150 that will show seasonal patterns and a lot we're known already like you may love anyone see, not the resolution we've done now. And then many were brand new that people didn't know before. So we have the seasonal molecules how many patterns do they fall into. And so you can actually show this mathematically you can cluster them. And it turns out, as I say mathematically you can show their their two patterns now these are people profiled in northern California three or from Southern California. You sell cool course that makes sense you're in California is only two biological season they should be winter and summer and there you would be half right. There is a winter strong pattern. That's this here. And so the and it turns out the molecules go up and winter late late winter December early January. It's things involved in illness, even acne this is known already this goes up in winter. And then, and for the things that the other pattern though is not summer it turns out in hindsight it makes sense as late April, early May. And it turns out that associates with asthma and allergies which again makes sense when things are pollinating. And then there's a lot of other things that in hindsight sort of makes sense some of it was known but a lot's not known cardiovascular disease, metabolic disease type two diabetes all these things are peaking in late April and the pathways associated with them early May and you may say why is that. And I would argue even in California people is not as active in winter it rains, you know it gets into the 50 that's that's cold for us we don't like that. So, the bottom line is that we think them people are you know less active in these months, you know not so good markers are piling up. And then spring comes and you can actually do that and there's other things I don't fully understand about seizures and sleep and stuff like that to that. And we think this information is clinically actionable because their laboratory tests clinical labs that associate with these seasonal patterns. And so you actually could take this into account in your healthcare system. And of course you never get this information just from genomics. I'm showing you the clinical version you get a much better resolution and use the whole money it's profiling from these patterns I told you earlier. Your microbes in both your nose and in your gut shift and I'm, I imagine, obviously the nose is no brainer because you're going outside and patterns are different, and your gut probably reflects the food you're eating. They're also different but it also turns out your gut, we can show this in other ways is very very in tune with your environmental exposures as well it's pretty cool paper and bio archive check it out. So the other time that we see shifts in people's molecular patterns changing is with time and we're calling that aging. And so we have enough people we think you need just five measurements within two years and we can tell how people are aging and the cool thing that one minute Mike. Okay, better hustle. Alright the cool thing we found is that everybody's aging differently so this is me I'm a pretty typical age or cardio. Age or this person's a sorry I'm a coagulation metabolic age or this person's a cardio age or so people are aging different patterns we think of this like a car where the car gets older. But some patterns are, you know, some things are wearing out fast and others and it turns out this person stage to hypertensive. We can group people in the patterns that we call agotype so these people each comes here these people over here they're aging in all these different categories. Here might be this guy's a kidney age or his kidney metabolic. Luckily I'm not much of an immune age or although age and the other things so so anyway we can tell people are aging their clinical markers associated with it and we can actually show that some people are actually reducing these markers and their their age of types and we in this case we have some interpretation as a lifestyle change other cases less clear like creatinine. Some people have actually shifted their their their molecules down and we can see our people aging based on these molecular phenotypes with Morgan Levine's pheno age kind of cool everybody's aging differently. I'll just make one last plug since came up on wearables they shuffled some slides at the last minute. We're very keen on this because it's these things are measuring 24 seven you may know 50 million people in the US where smartwatch 20% of the population the deepest watch makes hundreds thousand measurements every day. The expensive ones make 2.5 million. And so they'll measure high rate high rate variability they're all pretty good for that and then they're questionable in terms of absolute values for for blood oxygen blood pressure skin which is pretty good ones out there now on this. And so we've been using using to make a shameless plug we I actually think your smartwatch can be your most important health device in the future be or or some equivalent there because it does relay information back we now can detect the COVID early with an alerting system so making a shameless plug for you to sign a first study. If you have a smartwatch. This is a case where this person's wrestling heart rate jumped up nine and a half days before symptoms. It's four days by the way so we can alert and with the alerting system it's three days not it's complicated why we're not quite as sensitive but the bottom line is, we think for millions of people we can follow them a real time and alert them when their heart rate jumps up. When they get ill and for COVID you have a long pre symptomatic period so three to four days very very reasonable. And that's a big deal again then you'll stop people from spreading. We think it's going to be powerful pull all this data whether it's wearables clinical even your microbiome data we build a dashboard to be able to pull this in and follow it at any resolution you want. This was spent on quite a few years on this, by the way, predict other clinical markers I don't have time for this but including like your your blood hemoglobin levels and things like this it's not clinical grade values but it's good enough to give you hint something and maybe you can see early signs of anemia, also glucose etc. And finally this is a this is the future I'd like to see reflects everything you just heard in the earlier discussion. People's genome sequencing whether it's a rare disease or complex disease when combined with other sorts of measurements omics wearable data we think it'll be very very powerful for predicting disease risk and catching disease early. And then I have an amazing team of people who's worked with me over the years on this. So these are some of them I won't have time. And this is our wearables team we scaled this up rapidly in response to the pandemic pandemic to be able to try and have impact now. So that's the story happy to take questions I assume we have time. Yes, we have time so thank you for amazing data let me take the first question followed by Howard. You mentioned early in your talk quantifying stress and I'm increasingly convinced that this is relevant to many people on the call. So, can you expand on kind of quantifying stress and what kind of what kind of impact that has. Yeah, so believe it or not we can pick up stress was a smartwatch by increased resting heart rate. It also turns out one things your smartwatch certain smartwatches measure is your galvanic stress response which is conductance on your skin. And when you're stressed you usually sweat more. And so you can actually pick that up and the combination is quite powerful GSR plus heart rate. I'll tell you we can tell them people are well stress from heart rate by especially first thing in the morning grab somebody's resting heart rate you know their mental health or their physical health just from a heart rate measurement. So that that one's turns out a little easier thing now depression I think the best ways to get that you may know there's a lot of how people interface with their electronics like their email or their web searches. There's a probably good mental stress indicator for you know whether the depressed or not, based on the kinds of searches, just even their speaking rate and the words they use is going to be very very impactful for mental health. Howard. Mike thanks so much for sharing this amazing experience I have a meta question about kind of the maybe study design that is that these multiomic approaches obviously it's you get all kinds of information. Going forward do you think that's more productive to actually select for some sort of at risk population here towards a certain question, and you can really do sort of these kind of follow ups to try to demonstrate that benefits of early diagnosed problems, let's say diabetes versus kind of like an all comer approach, you find out everything but then of course you're having all these one off examples of different kinds of benefits, right but it's much easier to recruit in that way. What are your thoughts. I think we need both. So, totally conflicted on this out on my first slide but we've spun off a company called q bio, this doing a medical version of this so it doesn't do, for example, the microbiome or any see because that's hard to interface with medically, but by doing these deep profiles being agnostic going across a broad range of areas. Same thing they've uncovered and they've have whole body MRI which by the way the medical establishment hates, but they've got all wrong it's not having nodules it's are any of the nodules growing that's the key. But anyway, by doing these deep measurements on people same thing we found early discovery broadly across a wide range of areas because you quite frankly don't know what's wrong with you right now Howard if you have something wrong with you. So the broad survey at some minimal level is a big deal we would argue to get that longitudinally repeatedly would just be incredibly powerful if we could do that for the whole world. And smartwatch gives you some level that but on top of that way you said it's true to if you really want to dig in like we don't know all the best medical markers. So by going to at risk populations like cardiovascular disease, we can probably get better biomarkers so we make sure we feed these in to what will be a medical test if that makes sense so at the end of the medical establishment isn't going to deal with this deep that's a research project that the end they need clinical value measurements even if they're multi analyte and that's okay but I think so I think from a discovery standpoint I think drilling into at risk populations will help us a lot probably save a, a lot of bang for the buck but I think for general health, you want to get as broad profiling as possible because it's you want to know what's going on with you period even not just in the areas you think you're at risk for. I think we have time for one last question we have two type of questions about connections to the EHR, or is it just going to be direct to consumer. Well I think you should be able to pull in your EHR it's very hard it's easy to say it's hard to make that happen. But yeah I think it should go direct to consumer the problem is these days it has to go direct to consumer, because nobody pays you to stay healthy. So the most medical systems aren't set up that way they only pay you when you're ill, very rarely do people get measured when they're healthy so I think you're going to have to do it yourself and I think you're going to have to pay out of pocket for it right now I hope we hope to change that in the future. Thank you for really stimulating talk Michael. So now that's the end of session one so I'm going to hand it over. Again, Mike if you don't mind kind of typing there are some additional questions of people can type into the chat. I'll hand it over to Dave Bodine.