 Okay, hearing none and still on time, we're calling to the podium Rex Chisholm who I think probably holds the longest record of being the chair of the principal investigator's group of probably any consortium of any NIH group and for good reason. Rex and I were chatting on the way over that it really is a case that emerged I think from day one had this remarkable character among competing academic institutions of just friendship and fellowship and Rex has been in no small measure a contributor to that good feeling about good science so Rex you have the floor. Well thanks Dan, it's very kind of you to say and I am giving this presentation on behalf of all the Emerge Investigators both current and those that are alumni from the past it is indeed I think a remarkable consortium from the standpoint that it's really a great collection of people that I think have become friends and really have figured out how to work really well together and over the last ten years the number of conflicts we've had is so incredibly small that it makes it just really a pleasure to be part of this network. As you can see here from this map the list of participants includes nine clinical sites it includes two sequencing centers it includes a coordinating center I got to call it the coordinating center it's a great example of how having people organize and help keep us on track has been I think a major component of the success of Emerge over time. The way Emerge gets its work done is accomplished through a series of work groups a clinical annotation EHR integration genomics outcomes pharmacogenomics phenotyping and return of results LC and most of these have carried on throughout the course of the history of Emerge this is where the work all happens and one of the things I think that has been key to success of Emerge is that a we meet three times a year face to face that's unusual for consortia and I think it's one of the reasons that we've gotten to be so effective at working together with each other but also because we spend a fair amount of the time at these meetings in these work group sessions so the work groups actually are having plenty of face-to-face time in addition to even more plenty of telephone conference time. We've also been able to benefit from the fact that NHGRI has been generous in providing some additional supplements geocoding is an important one health care provider survey is an important one I think geocoding is an exciting opportunity because it starts to let us think about how to integrate the issues related to environment into the gene by environment interactions which is probably one of the most difficult areas and I think one where Emerge will be able to make some important contributions and then recently we were awarded another for phenotyping to get emerge all across the Emerge sites aligned to the OMOP data model which allows us to begin to think about how do we increase the efficiency of data sharing between all of the Emerge sites and I think this is going to be an important activity going forward and we're going to see an acceleration the pace of our ability to do phenotyping going forward and then there are some subgroups for Emerge that also do the work similar to the each of the groups that we've seen as Dan alluded to in Emerge one there was skepticism about how good data for electronic health records was so in Emerge one one of the big questions was can electronic medical records in Biobank samples be used for genomic research and genomic discovery we focused on genome-wide genotyping and GWAS studies and I think that was a very successful process and by the end of Emerge one we had definitely confirmed that electronic health record data actually could be used very effectively for research and it also established the fact that not only could it be used effectively but it could be used in a reusable way and so one of the things that I think was really important about Emerge one is to realize that once we had genotype samples for one purpose they could actually be reused and refenotyped over and over again to do genomic discovery with different phenotypes based on having genotype information so it's a very efficient approach Emerge two which ran from 11 to 15 focused beginning on thinking about how do we do clinical implementation how do we begin to actually put this information back in electronic health records and use it to improve our health records our genomic medicine work and this part of Emerge two we also picked up the pharmacogenomics project where we actually genotyped actually sequenced the genes using the PGRN C platform for the 84 genes that were important for a drug metabolism and actually used that information in a way that was put back into the electronic health record and used to help predict what kinds of medications people should be given and that project is actually continuing on to this day and then Emerge three we moved into the world of sequencing we began to think about sequencing a panel of genes we looked at a series of 109 genes which include the ACM gene now 59 genes thought about how do we do clinical interpretation based on genetic variants that are found in those sequences and as Emerge has done during all of its sections it didn't stop doing anything it just kept adding stuff on and so we continue to do GWAS and electronic phenotyping in Emerge three as well so the specific aims of Emerge three are to assess sequence and assess clinically relevant genes and we're gonna do that in a cohort of 25,000 individuals we're actually about over halfway there already assess the phenotypic implications of these variants integrate the genetic variants back into the electronic health records and to think about how do we use that for care and I think that's one of the unique features of Emerge is the systematic approach of delivery of electronic of genomic information through the electronic health record and then we have throughout history of Emerge I use this as an opportunity to create community resources that can be shared not only with the Emerge network but with the research community more broadly so where are we today well so right now we have in Emerge over 110,000 individual genomic data set consists of both the data on 100,000 I'll show a little bit more about this in a minute it includes an Emerge record counter which allows us to go in publicly available site you can go in and ask how many people meet a set of criteria that I'm interested in type 2 diabetics with the BMI over 40 and get a number back so get a sense is there a feasible opportunity in this data set to actually do some kind of research study going forward and then Sphinx which came out initially of the pharmacogenomics project but which allows us to look at genes by drug or other kinds of act pathways or other ways of looking at these genes and all of these have been a pet we paid attention to ancestry and a focus on diversity and so you can see that we're right now at around 111,000 participants we have some additional samples coming in so by the end of Emerge 3 we imagine we'll be up to about 135,000 different individuals covered one of the big tasks has been to deliver an imputation of the GWAS data what you can see here is a principal components analysis of this initial data set this is about 84,000 individuals what you can see is that it really does represent quite a diverse data set it's especially valuable for thinking about how do we look across a variety of ancestries for discovery purposes another deliverable has been the development of the Emerge Seek platform this is this platform of 109 genes that we're focused on it included the original ACMG 56 we went through a process where each site was able to nominate its top six genes based on the kinds of research it wanted to do locally those were proposed to move forward and then after some discussion and refinement with the group in general we end up with the 109 proposed genes after a great deal of work by our clinical annotation group we came up that 68 of these are actually clinically actionable and then the remainder are there for discovery purposes going forward there's also a variety of single nucleotide variants that are on the chip and again there are 14 of those that are clinically actionable and so of those there are 68 genes and 14 SNPs that will be returned by all sites during the course of this project so we are in the midst of the process of returning the results that we have and beginning to learn I think important new results that will get important new features that will get about how do you return results how do people react to the return of results and are there lessons learned that we should think about going forward as we think about implementation of genomic medicine to date almost 15,000 samples have been sequenced with 3700 reports issued and those are in various states of being returned to participants back at each of the clinical sites just to give you a sense this is from a few weeks ago but you can see this is the partners interpretation what you can see is the vast majority of the results are negative regardless of whether you look at indication based or non-indication based return of results the number of positives results in a range from two to three or four percent that's pretty consistent with what people have seen in other studies as well a similar results are seen from the Baylor sequencing site and I think one of the things that is important to remember is that one of the features of a merge by actually having two sequencing centers is a little mini-test of how genomic medicine might actually have to play out with multiple research lab or clinical laboratories reporting back and it I think has been very instructive to us to learn the differences between the two clinical laboratories that we've worked with and how do we think about integrating the data from different laboratories together this has been a work a great deal of work gone into by the clinical act the clinical annotation work group as well as our EHR working group to think about how to how to deal with some of these things phenotyping is another area where we think we've had a huge impact certainly figuring out how to do clinical phenotyping based on electronic health records data has been a challenge it began in emerge one with us starting to think about how does one go about developing algorithms and actually just since it's not obvious to everybody you can see here's just one example of an algorithm for type 2 diabetes not to go through it in detail but what you can see is that it consists of multiple data elements multiple branch points so it's not sufficient to just look in the electronic health record and ask the question is there a diagnosis code for type 2 diabetes you actually have to go through if you ask that question directly you'll get a very poor outcome but it is possible using an algorithm such as this and refinement to get to 95% or better positive predictive value and so what we've focused on is then creating phenotyping database that you call VKB which collects all this information together there's a public facing side where everybody can see what the details of any given algorithm that's been published is and then there's a private side that emerge uses to work on refining its data phenotyping algorithms I think this is demonstrated the feasibility of the use of electronic health records and phenotyping in genomic medicine we've created a variety of tools that will spend a little bit more time in a couple minutes talking about the way the workflow happens in terms of phenotype development begins with one site picking being the champion for that phenotype development they create an algorithm and then once they think they've got that algorithm working at their site then we move into a validation phase and a sharing phase where it moves on to another site the second site then runs that algorithm and inevitably in the process of this iterative process of additional sites running the algorithm we learn about specific things that one site might have that causes an adjustment to the algorithm need to be made and the algorithms are obviously changed iteratively over the time of this process of the workflow once we feel like it's met a sufficient threshold of having appropriate positive predictive value we then move on to publish the phenotypes algorithms so that they're then deployed across the entire emerging network by all of the sites and then that's used for the discovery purposes that we've talked about a little bit and all of this is managed through this process of feet KB using things like the record counter that I've already alluded to and data dictionary and data set validation moving to the OMOP model I think will actually greatly increase the efficiency of this and I'm hoping that in subsequent phenotyping efforts we'll see that this iterative process is simplified because every site will now have a similar data modem and be able to use that information in a more efficient way in emerge one we focused on 14 total phenotypes in emerge two we added an additional 15 phenotypes that brought us up to a total of 29 and finally in emerge three we're going to do a total of 27 phenotypes so the total of 70 phenotypes over the course of the history of emerge will be deployed by the end of this coming summer one I think really important contribution that emerges made and really need to point out the leadership of Vanderbilt and Josh Danny in particular in terms of thinking about how do we take the GWAS approach and really flip it on its head and I think emerge has really been a key in the development of this idea of a phenome wide association study where basically what you do is instead of asking the question for a particular phenotype what genotypes are associated with that phenotype what we take is ask the question for a specific genotype what phenotypes are associated with that and that phenome wide association study I think has really revolutionized the way a lot of people think about how do you think about the phenome how do you think about associations between genetic variants and and phenotypes and so as a demonstration of this the emerge groups looked across all of the NHGRI GWAS catalog SNPs there were three hundred thirty one hundred SNPs that were associated with prior GWAS associations and there was a both a replication arm which was able to show that we were able to replicate 751 or so associations and a discovery arm that was able to actually replicate produce not replicate novel associations that have been made as well so this process of looking at it from a phenome wide approach was really able to help us think about the fact that you could actually reproduce what had been seen in the genomics GWAS catalog and so I think it's a really important way that emerge has contributed a different way of thinking and I think these are being widely used and certainly emerge continues to use GWAS approaches in many of its projects going forward the emerge pharmacogenomics project I mentioned a little earlier consists of about 9,000 participant data set with sequence of 84 genes that were important for drug metabolism there are still the data is been cleaned and reanalyzed and sites are now still in the process collecting utilization and outcomes data on whether or not this makes a difference in terms of prescription medic of medications to our participants the next deliverable is the PGRN seat multi-sample calling it's a little bit smaller data set again you can see the principal components analysis of the 9,000 data set here the diversity just because it's a smaller sample is there but not as well filled in as it was in the hundred and ten thousand or the eight I guess eighty four thousand in the last picture but you can see we still have pretty good discovery across that group another area where I think emerges had huge impact is in the area of genomic data into the electronic health records there's been an enormous amount of work in infrastructure put into tools and clinically clinical decision support tools thinking about how do we put genomic variants back into the electronic health record there's been a lot of work about ancillary genomic data systems which think about how do we store sequence data without coming up the works of an electronic health record but with the ability to efficiently put data from a genome-wide association study an exome or even whole genome sequence and think about how do we integrate that back into electronic health records a lot of this has been led by also work in the info button project how do we use existing resources embedding them into electronic health records in a way so that they can be widely used by a wide number of health care providers and this is an important feature of emerge because it's enabled the product of emerge to be deployed pretty easily in many electronic health record systems including its sites that were so sophisticated and involved actively in emerge and then another deliverable that we have is the clinical decision support knowledge base which is where we in partnership with the ignite network are creating a catalog of clinical decision support that's been put in place and enabled I think again a tool that can be used by other sites as they think about how to roll out genomic medicine at one of the other interesting features that were in the midst of right now is thinking about using network wide analysis using a cloud platform DNA Nexus this is an opportunity for them for multiple sites to create common tools that can be applied across the entire data sets we're in the midst of figuring out how to do this right now but this data will be in the cloud will be our GWAS studies the PGRN seek and all of the emerge seek data as well as some whole genome sequence data that we were able to achieve with some help from some of the genome sequencing centers and so I think as we begin to think about this platform in the cloud this is a novel way for us to begin to do genomic analysis going forward and again has the opportunity create some real efficiencies so that every site doesn't have to create their own data sets and their own pipelines but rather can use a common infrastructure in the cloud but it also doesn't prohibit ability of individual sites to experiment because you can create apps based on this infrastructure that allow you to experiment and pursue things in a novel way so I think a cloud-based platform but allows innovation to continue in a way that I think it will be a real interesting demonstration project going forward for how do we think about some of these activities merge has been very productive over the course of its time what you can see here is the publication record there are over 633 total projects between manuscripts published in manuscript concept sheets that have that are currently in the works you can see here that in the blue site specific projects and in the gray that's gray the projects that are site-wide are ESP external scientific panel and our chair Howard McLeod have constantly reminded us that we need to be working hard to maximize the activity so you can see there's been an enormous growth in system network-wide projects over the course of time we're very proud of the fact that not only are these publications produced but they're actually being widely cited by the research community and as of March of last year over 17,000 citations for emerge it's really quite an impressive contribution to scientific literature and it's one that's clearly having a huge impact you can see a lot of it is in the area of genomics so very appropriate for the Human Genome Research Institute that this continues to be an important area for contribution from emerge phenotyping continues to be an important part of this return of results and Elsie kinds of things have also been something that's proceeded throughout the course of emerge and I think it continues to be a very important activity that emerge participates in and then a variety of other categories as well one of the other ways to measure the impact of emerge I think is to ask the question all of our data is of course deposited into dbGaP is anybody else finding it useful by downloading it from dbGaP and what you can see here is that phase one phase two and phase three data is being used pretty widely obviously phase one has been in there the longest so it's being used the most but phase pre phase two means the folks that were bringing new data into phase two of emerge and you can see the data continues to grow we're about to launch mid-term dbGaP submission of all of the emerge seek data and so that will be yet another dataset that would be broadly available to the research community and so you can see we also get a lot of data visiting our website you know 1500 sessions per month over a thousand users per month VKB website has seen a lot of use average in the last six months 50% of them are new people coming into the system views from over 76 countries around the world so it's not only having an impact locally but an impact internationally I mentioned that the tools development was an important component of what we think it was that emerged set out to do I've talked about many of these the VKB is this knowledge base for phenotypes my results org is a tool that is available that talks about how do you interpret your results and it's a very useful tool to be able to have electronic health records information link out directly to a site that somebody can actually go look a participant could actually go look and understand what their results might mean Sphinx I've already talked about info button I've talked about one of the things early on emerge did was created model consent language that I think has been pretty widely adopted for how do we think about data sharing and how do we think about consenting and then the fee was catalog that I've also talked a little bit genotyping tools clinical decision support tools phenotyping tools and natural language processing tools are all part of the armamentarium we have to actually be able to do the work of merge and all of these tools are widely shared widely available to the entire research community so where are we now halfway through or a little bit more than halfway through emerge three we have some deliverables db gap submission I mentioned that an interim submission is in the works right now on about half of the emerge sequence data we are we've established it support for return of results purposes and again I think thinking about a model for what emerge might be able to contribute thinking about two different clinical laboratories reporting think about multiple EHR systems that they're reporting into we are beginning to converge on some models for how you might do this efficiently going forward that could actually be adopted widely I think at other sites as well and maybe even broadly across people when she to implement genomic medicine as I said we've shared clinical decision support through the cds knowledge base and we're starting to look at outcomes for the return of results and it's early days we're still in an anecdotal state on return of results I think one of the things that we're learning is that things that we might not even imagine are clinically actionable people are finding of value as we start to think about returning results to them and so I think this is going to be an important area for us to start to look at each site is coming up with some examples of this and you know in one case I'll just mention from our place you know we've got some variants that are linked to obesity and participants that have that condition you know I think it'll be interesting to learn from them what the consequences of and the value to them of returning that results even though it may not be specifically clinically actionable you could imagine it might have an impact on them and their families so we're also continuing to deploy phenotypes we are phenotyping work group is really got the pedal to the metal and the whip out to make sure that everybody's completing their phenotype works and we get them deployed as well so I think I'll stop there I realize it's been a whirlwind tour but just to sort of close by saying in summary then as if this you know needed this whirlwind tour needed summary what has emerged accomplished well I think first emerged has demonstrated that there's real value to the HR of the HR for genomic research that was an open question we've heard about that at the beginning I think that question has been firmly and conclusively answered and creates real opportunity for us thinking about going forward and at the risk of maybe offending some I think the fact that emerge was able to demonstrate that I think in some small way at least contributed to the willingness and maybe even the support for thinking about rolling out something like the all of us research program because it makes it possible to think about how to do this broadly across a very large cohort in the United States I think emerge also created a genomic resource which we can highly phenotype repetitively over and over and over again the efficiency of that cannot be underestimated it's really incredible think about the all the effort that went into build purpose-built cohorts in the past we now don't have to build purpose-built cohorts we can just take data from a large collection and reanalyze it in an appropriate way that was an important I think contribution emerge made I think we created an important test bed for genomic medicine both research and implementation I think that has impacts in terms of how do we return results how do we get that into the electronic health record how do clinical laboratories provide it to health records one of the standards the data standards that should be met for that and I think emerge has got some great ideas about all of those areas where going forward I think other places can benefit from what we've learned it emerge we we've developed the concept of SLR genomic systems and as I said data standards for returning results we've created a library of phenotyping algorithms and I could go on and on but I think it suffice it to say that hopefully I've given you enough of an overview of what emerge is accomplished in the last decade of work that we can have that as a base for a robust discussion today about what the exciting opportunities are going forward so I'll stop there and happy to take any questions any of you might have questions for Rex Dan's going to give me a softball question there might set up in the front what remains to be done and and at what point do you hand this off to a larger right so it's a it's a great it's a great question so I think there is a tremendous amount to be done we still I think have a lot of work to do in terms of understanding and building the infrastructure for returning results into electronic health records we have a lot of work to be done in terms of the infrastructure of how do you want to get it into an electronic health record give it to our participants and their providers in a useful way I think we have a lot of work to be done in terms of generating evidence for the value of whether it's pharmacogenomics and returning results or whether it's other gene susceptibility increased screening a lot of work to be done in that area still a lot of evidence we're still I would say in very early days of evidence generation I guess I also would like to think about how emerge relates to programs like all of us first to say many of us in the room are very active participants in all of us as well and so I think we like to think that we bring what we've learned from emerge to the broader all of us community I also like to think of us as I think of the great story of the defeat of the Spanish Armada by the British emerge maybe the nimble small ships that can turn on a much shorter time frame then maybe the large aircraft carrier of the all of us project terrific project but we might be able to be a vanguard that can help set the stage for how to think about all of us as it goes forward because it's got a long history a long future ahead of it and I could go on but maybe that's good enough five or six reasons that we think there's still much more work to be done so Eric so you use the word resource and genomic resource several times in fact I would claim NHGRI is at its best when it creates large resources for the scientific community to further mine but one of the things you didn't mention is moving the I think 70 phenotypes and 135,000 individuals how are those data moved into db gap and how are they made available to the scientific community well they moved in the db gap in the standard way deposit you know going through the deposit process the availability of phenotypes to the extent that you need you know individual level access of course it's subject to all the db gap restrictions policies but yes yes if that's what you're asking these short answers yes but let me also use that as an opportunity to say one other thing about emerge so emerge has been very active in seeking partnerships with people outside of emerge we have several affiliate members in the merge who can actually access the data in a way different from db gap and so there might be value there to people thinking about partnering with emerge as well so although the data is in db gap there may be more efficient ways to get it and one of the ways is to become a merge affiliate so I you described the 70 complex phenotypes that have been done it is there a library or a central repository of other variables I mean I think the word phenotypes can be applied narrowly or more broadly things like demographic variables much simple much more simply cleaned elements and are those aggregated as well well yeah yes those are and you could either get at them through the record counter kind of in the record counter that I alluded to will give you some of that information and then of course it's also that some of that kind of information is available in db gap but we're always happy to do queries if that would be helpful to provide additional information but basic demographics I think are pretty readily available from from the record counter um so Rex I have a question about kind of sustainability of tools so like I'm looking at my results org and of course I look at the syndromes I know about and and I would say I don't necessarily agree with the recommendations they're on here but I was just curious is there some systematic way in which these tools or someone's job it is to look at them once a year I mean because whenever you create something like this it sits out in the web forever and so I was just curious how does a merge look at all of the various tools it's created and updating them so it's a it's a good point it's it's actually a real challenge to keep all of those tools up to date as you as you know the other problem is there's a challenge in terms of so the tools are built for research purposes I think by and large the tools that emerges created I'll use record counter as an example I use things as an example I think they're robust but they're dependent on people continuing to maintain them and so you know I think one of the questions to think about going forward is if there is value to those tools we should be thinking as part of our discussion today about how do we assure that those tools are sustained and maintained going forward and I would say not only sustained and maintained but actually upgraded as appropriate yeah this builds on what Eric and Sharon had asked I mean I think the point is well made about the sustainability of this but I think it's also important to recognize that one of the things that emerges done well in my view has been to partner with other groups to try and make sure that we can it's not just falling on us to do it so for example we work very closely with ClinGen to make sure that things that we're producing that are relevant to ClinGen can be incorporated with within their model which I think has a little bit more juice behind it in terms of sustainability we're also working with the ignite group in terms of creating an outcomes tool kit that can be available although both of us both networks are still under you know a grant funded mechanism so there will need to be a sustainability plan but it's not solely on on eMERGE and then the other thing to mention is it's not just the data in my view that eMERGE has been contributing but it's also been contribution to methodologies things like FeeKB where the phenotyping algorithms are out there and available for anybody to use and I know we have a record counter that shows how many times those have been downloaded and so the impact related to these types of activities also I think is in the positive ledger for the network. Steve Leader, Children's Mercy Hospital in Kansas City. I was wondering to what extent you think that the the developments and the infrastructure that the accomplishments that have been made so far are now ready for moving into the pediatric space. I'm thinking specifically of things like neurodevelopmental disorders, maybe inflammatory diseases that have their onset due to neolithic arthritis, inflammatory bowel diseases. Well so I guess the first thing I should say is that eMERGE has since eMERGE too had two pediatric sites since a children's hospital and CHOP so they've been very actively involved in working in pediatric space. I don't know John do you want to take a crack at answering that? You need to come up to a microphone. John Harley PI at the Cincinnati Children's. I think the integration with pediatrics and adults is a challenge but however the there are many phenotypes that span both pediatrics and adults and of course the infrastructure issues are about the same for everybody and there's a lot about implementation that eMERGE does that'll make the whole system work better. Thank you I'm Jesse Tenenbaum. I'm from Duke and I'm an outside observer so I'm not as familiar. I thought I understood based on what Terry was saying the distinction between CSER and eMERGE but the way you described eMERGE two and three it seems like they're converging a little. Can you either confirm or deny? I can confirm. I'm not going to get subpoena for the yeah I can confirm. I mean so in fact one of the things that I think again NHGR can be very proud of is the fact that CSER and eMERGE have had at least two joint meetings together. Many of their work groups continue to work together. I think the difference is that CSER at least initially was focused on individual participants with a particular condition whereas eMERGE took this very much broader view. So to say that they've converged is probably fair because I think as we've thought about eMERGE three where we're sequencing 25,000 individuals again not necessarily with respect to a condition because over half of our participants in that 25,000 are not being sequenced for an indication whereas in CSER the focus has been mostly on indication but yes some convergence for sure. Howard did you Yeah I wonder if you could comment a little bit about how the impact you've had on going across the different electronic medical record vendors if you will. You know even within a given brand there are you know every one of them seems to be different so if you'd comment on that part. Yeah there's a saying amongst electronic health records folks that for those many of us who have that epic installations you've seen one epic you've seen one epic. So I think we've learned a lot of lessons about how to go across again I'm really excited about the OMOP model and how by rolling that out across our sites we'll be able to think about that. You can think about how that might be a model going forward for health systems broadly is to use those kinds of standards. I think one of the things that we've learned from day one of eMERGE was all about trying to deploy standards and using standards and so I think we've really figured out how to do that. It is still not the case that an algorithm at one place is plug and play across another place. We've done some experimentation about that. We've actually created packages of computable phenotypes that could be shared across sites and they're definitely not plug and play but I think maybe with OMOP together with use of some other standards we might get there but I think what we have learned is how do we extract common data elements from different electronic health records and how do we learn about the variables or variables of different sites and how they use that electronic health record in different ways to do different things and we've figured out how to overcome that so I think the lessons have been pretty strong in terms of thinking about how to bring data together from multiple healthcare systems and integrate them into a single hall. Yeah just to add on to that as a former chair of the electronic health record integration working group I think that again we we sometimes focus on the front end of the implementations which is of less interest to eMERGE than the back end which is where the data is actually being produced and I think while there are differences there there are considerably fewer differences there than at the actual interface with the clinician. I think the other point to make though is that now that the vendors are beginning to understand that they need to be incorporating genomics into their systems they look to eMERGE as being a leader in terms of how to do this and so many of us are engaged with the vendor community to do this and one of the other initiatives that we participate in that wasn't specifically mentioned because it's not an NHGRI activity although NHGRI is involved in it is the digitized collaboration which came out of the at that time the Institute of Medicine where there are research clinical and vendor-based representatives around the table that are creating actual implementations that are being rolled out and tested in the real world. Two pharmacogenomic use cases have already been implemented through that there's one that's now being looked at related to Lynch syndrome so I think that the influence that eMERGE has had in terms of telling vendors this is what you really need to do and here's the expertise about how to do it has been extremely influential. Can you go to microphone? Along these same lines is there somewhere for those of us that are struggling with our own institutional electronic medical record uh is there a menu can we go somewhere to see a menu of what has been worked out at all of the eMERGE sites so that we can say look this epic has managed to do this why can't we do this those sorts of things. Yeah I don't know that we have a catalog Sandy you want to comment on that? Yeah there there isn't so much a catalog I mean what to extend on Mark's point what's happening is the vast majority of the sites wind up needing to do ancillary genetic systems in order to the support I think epic has been great in the digitized context in terms of expressing interest in working with us to try to improve support but the support in epic right now is just so primitive that that there's a need to do things outside of epic but again to extend on Mark's point I think there's an enormous opportunity for us to to demonstrate how these things need to be handled so that epic becomes more aware of how to build that support. Go ahead Casey. So I was just going to add that there's also CDSKB where people are starting to share their less you know artifacts related to decision support that they're implementing locally and with the right approvals then they're able to also show the interface of decision support and post those and make those publicly accessible and also docuBuild is another resource that we're working on an e-merge and that is allowing groups to share content and customize that content wherever their local needs are and so that's something that we're just beginning to roll out if I could just follow I think all of that's great I'm wondering to the naive sort of person to emerge where does one find all that is it is there a website where one can go to see all of those relevant yeah there is a there's a tools website but I don't think it's gonna it's not going to achieve what you want and I think it's a really important point and it's something that a merge should discuss I think in the maybe get the HRI group thinking about this is how do we better display all the lessons that we've learned because we're deep in drinking from a fire hose in terms of thinking about how to how to do that and how to receive now Richard standing up at the microphone reminds me and it's just the amount of work that's gone into understanding how to receive data electronically from a clinical laboratory rather than in the context of what they like to do which is a pdf has been a effort that we've put an enormous amount of time into and the clinical sites working together with the sequencing centers and clinical labs I think are really struggling and succeeding with that well thanks Rex I thought I should just mention that David's question reflects the importance of this issue I think that's on everybody's minds the two publications that are being worked on right now Sandy and Casey have worked on one fairly targeted one that's been submitted as a more general one to describe the experiences and I think the amount of work that's had to go into this question and answering it at individual sites is going to precipitate the same kind of big descriptions to them at the risk of being perhaps a little bit facetious this could represent a sustainability model if we develop a consultation service available to people to help them implement this so but it's just something to consider probably outside the NIH model but I start up company in the making yeah yeah exactly so but I but I do think that we we've attempted through traditional methods of dissemination to get the information out there but I think it's clear that that's been less than successful and so it's a it's an interesting idea to consider in the last year of last year plus of funding about how we might be able to look at this more from an operational perspective this is something that we've done in ClinGen again in conjunction with Emerge to look at the infobutton project was mentioned which is something that any certified electronic health record has infobutton capability and whether your system has turned it on or not is not clear but if they have turned it on we can deploy a virtual machine to your site that will essentially upload the ability to consume ClinGen resources using that infobutton and so we've made it easy to implement and the more we can do of those sorts of things I think in the long run the bigger impact that this will have we have to put air quotes around easy I think yeah easy as is anything easy in the EHR space of which nothing is easy I would just advocate adding a psychotherapy module to that for the emergent investigators um so onto a slightly less headache producing topic I'm not sure I understand Sphinx versus FarmGKB and CPIC and could you just again it sounds like these are somewhat overlapping reasons yeah Sphinx is a repository of Emerge data so that's what Sphinx is so it's not meant to be the breadth of FarmGKB it's really meant to be a way that people can come and look at Emerge data Emerge data with regard to the pharmacogenomics sequencing yeah of those 84 genes including allele frequencies and I don't know if this is microphone on but David was saying including allele frequencies amongst our cohort but it's not meant to be a knowledge base about those no okay thanks it's a way you can go in and search the Emerge data set