 So I apologize in advance I've had some internet connectivity issues, not to mention issues with pediatric members of my house, but I haven't been fully on all day, but I did, I was asked to react in this area, so I'll talk about three different topics that I would ask to talk about, actionability, validation, and standards for lab reports, and just sort of go through them one by one. Next slide please. Next slide. Oh, thanks. I don't think I need to define actionability, but in my observation as an outsider to emerge, it seems there's been a large focus to date on genotypes, either pharmacogenomics, GWAS or FIWAS, and one suggestion is, you know, and I think it's been talked about, is to expand the focus to gene-phenotype relationships, not just relationships of genotypes. And although I don't think I have very specific messages within the area of actionability, I do think it's an opportunity for collaboration, and I thought I'd mention a project that's ongoing. Next slide please. So this has a few different arms to it, one through the baby seek study, which is one of the newborn screening genome grants. We've been doing an evidence-based review of gene-phenotype associations. We're through 673 genes and have 3,000 more to go, and we've been systematically evaluating the evidence for the Gene Disease Association in a ranking of minus one to three, as well as categorizing a variety of other fields on each gene. If we skip to the next slide, this project is being integrated into a larger effort that we're doing collaboratively with two other labs that has been focused to a large extent on defining the medical exome and improving testing for all genes that have known medical relevance, and that's been led by Emory Chopp and my lab. However, the evaluation of the evidence on variants, similar to what I just showed in the prior slide, is being extended to a larger program called the ClinGen resource, and on the next slide it defines the four different groups that are part of that effort, really creating centralized resources. And one of those groups includes a grant by Jonathan Berg, who is really focused on actionability as a component of this project and defining that for genes. So on the next slide is just a slide that I actually took from one of his slide decks, really identifying five key parameters that he has developed for evaluating medical actionability, and each one of those is scored on a scale from zero to three, and then he's working with Katrina Goddard's group to really generate a streamlined evidence-based review of all of these genes. And so that's being built off of the medical exome list for all of these genes, and we're doing that collaboratively. So I think that's an opportunity for collaboration among the many different consortia that may be working in this space, and I'm happy to talk more about it later. But if you go to the next slide, I'll switch over to the validation. So I thought I would bring it to people's attention that we recently published guidelines through the American College of Medical Genetics that addresses clinical laboratory standards for next-generation sequencing, and that includes how to validate a sequencing test on next-generation sequencing platforms, involving both platform validation as well as individual test validation and ongoing quality assurance. So if you skip to the next slide, that covers many of the detailed aspects of this and I think gives reasonably good guidance. I did want to point out a couple of specific recommendations that we made. One is that when you validate a test that you must cover all types of rare variants that are being reported, as well as address homologous regions of which there are hundreds of genes that have issues, either pseudo genes or other smaller regions of homology. And so your validation has to address the types of things you will encounter in your test and are planning to report. That is distinct from common variants that are, you know, for instance, pharmacogenomic variants where you're specifically trying to report out on a particular known genotype. And in that case, the validation would need to be variant-specific, so we make that distinction. There's also a lot of discussion about whether there's a need to orthogonally validate every finding from next-generation sequencing, and we acknowledged in the recommendations that may not be necessary, although it's fairly standard today, but may not be necessary if sufficient validation has been performed for, as stated above. And the quality metrics are high, both coverage and mapping quality, and most importantly, that the workflow has a low risk for sample swaps so that you can ensure that the result you got from your primary platform is, in fact, derived from the original specimen. And so some labs put in other methods, SNP arrays, and things to ensure the final results came from the original specimen. Next slide. I also wanted to just point out a couple of things about variant calling. A lot of people use traditional type lines to perform alignment in variant calling, such as JTK, that generate a variant call file. However, when you're looking at specific genotypes, there are a number of groups that are taking slightly different algorithms to analyze that data in a more accurate manner, and that could either be done through joint calling, where many cases, many data sets are batched. The problem with this is it's not a great approach for the rapid turnaround time that happens in clinical labs. So in our group, we've implemented individual genotyping calling on the raw NGS data for all of our pharmacogenomic variants and our complex common variants for complex traits. And that has been more amenable to our clinical workflow, and it also ensures return when there's a reference call and not just when the variant is there. Next slide. So the last topic is related to standards for lab reports and EHR deposition, and we did address a number of guidance in the ACMG document as well, including all of the sort of topics listed here, chair on times, interpretation, incidental findings. We're in report, we give a number of sample reports for both NGS panels and exome and address data reanalysis. I will point out one specific recommendation related to the EHR, and that was that variants that are going into the EHR today be restricted to those with analytic and clinical validity, and that those variants be structured to allow for clinical decision support, so the variants may be in structured form, and the interpretive part of it could be a PDF. And we did not recommend today that an entire variant call file from an exomer genome be deposited in the EHR until improved analytic validity. Next slide. This just shows an example of a sample report that shows the structured data as well as an overall result that we recommended be put on to individual reports. Next slide. And the other question that a lot of people talk about is how you return updated information on variants that happens over time inevitably, and so we, in our, in my lab, we specifically implemented a system called the Gene Insight Clinic, where this is a system that in our instance is integrated to the partner's health care EHR environment. In other instances where it's not integrated, it's just web accessible, but it serves as a repository for all of the clinics, genetic test reports, and all of the data in there, if you click the next slide, it will show a bump out, so when you just click next, then a variant classification changes. It will show in the system that then both, oops, go back one, the variant's original classification as well as a new classification, and physicians are sent an email alert with a link deep into the EHR record where they can find the information on this, and they can click a hyperlink to the variant, and it will show the updated interpretation of that variant based on connectivity with our lab's mutation database, so that's the way that we have implemented that sort of ongoing updating of variants, both rare and common throughout our genetic test reports, and this whole system is in the subject of an NIH grant that David Bates has led to look at usability by clinicians using the system. Next slide. And I just wanted to show a couple of examples of our whole genome reports where to really allow them to be usable to a primary care physician. We summarize all of the information on a single page, shown here is monogenic, disease risk, carrier status, pharmacogenomic associations, and blood groups, and then physicians can skip to the next pages, next slide please, and they'll see more details about each of the variants and the diseases being reported. Next slide shows one of the pharmacogenomic associations where they see the more detailed information as well as highlighted the background population rate of the genotype that they have compared to others. Next slide please. Although this isn't being reported in our routine clinical service for the MedSeq study, we are also adding a cardiac risk supplement that gets into complex traits. Next slide, and it will zoom in to the bottom section where we are reporting genotypes from complex traits, both contextual data as well as off to the right, their polygenic relative risk and percentile rank compared to the general population, and we've spent a lot of time trying to figure out how to return this type of data in a clinical report. So next slide please. Lastly, one of the biggest challenges we encounter is the interpretation of a particularly rare variant. Many of you know that there's ACMG recommendations out listing 56 genes to return that has led to a number of groups that have looked at the rate of internal findings in cohorts. Those rates have varied across studies, although I would say that many are around the 1-2 percent range for the 56 genes, but some of the differences come to differences in how we all interpret variants. And the next slide, when we interpret variants in a genome, we routinely find many pathogenic variants, 97 percent of which we exclude for lack of evidence, as well as loss of function variants, 94 percent of which we exclude because there's insufficient evidence for a gene disease association. So that presents huge challenges in the interpretive process, it's very labor intensive. Next slide, I think that might be my last slide, so I just wanted to acknowledge a number of the projects that are contributing to the things that I talked about today, and I'll stop there.