 Boston Children's, Milwaukee, and Philadelphia have, so they're four children's hospitals, I think represented one way or another in the group five. Who'd I miss? Why is Children's Hospital? Janet Wise Children's Hospital. Okay, and probably others. And so it's a different world for us. The, my CEO is connecting with other children's hospitals in fear that the new ration healthcare system may not leave a place for specialized children's hospitals like we've had in the past. And of course, I think that center picture is one of Mark at a younger age or a clone on the other. Ready to run a meeting. And so we see a lot of children. This is our city here on the left and our Children's Hospital on the right, nestled in the middle eastern, middle part of the eastern part of the country. We have about 1.1 million patient visits and have seen 70,000 new patients a year. We have about 13 outreach outpatient clinics. We have 5,000 patients in our neonatal intensive care units spread around the various sites. Locally we do 33,000 surgeries. We've done 500 liver transplants. So we do 3,500 adenoidectomies a year. We have some complicated surgical procedures that people come from all over the world for. And thanks to a endowment from the family, the Proctor family of Proctor and Gamble which is located in Cincinnati. Instead of building a building, an endowment was started and that has grown to about 1.5, 1.8 billion dollars for Children's Hospital which has made a big difference. We have an annual income from billing and everything else of about 1.4 billion dollars and we have an NIH income about 107 million a year. And we have not fallen off the physical cliff yet but I expect like everybody else it's just a moment before the physical discipline hits everyone. We have an epic electronic medical record and a CERNR lab record and the institution has the sense that we are behind and maybe kind of typical for a large referral center that we have not brought the genomic diagnostics and technologies to our patient care environment. As a consequence, I and the chief of human genetics, Greg Grabowski have been assigned the duty of trying to help generate a plan for the leadership that will help accelerate our progress. So at the moment we do individual gene sequencing, we do a lot of infectious disease work, we have a very nice cytogenetics lab do about 2,000 cytogenetic studies a year and have an active pharmacogenomics program that I'll talk about. We are converting to targeted gene sequencing and trying to implement whole genome sequencing and we are bracing ourselves for diagnostic testing and evaluation that brings to the table gene expression, DNA sensitivity, whole genome sequencing, methylation, histone marks and the chromosome immunoprecipitation testing. And I'm surprised at this meeting that there's been no discussion of these new technologies to speak of and we're completely focused on the DNA sequencing results and these I expect that if you think the world is complicated now, wait until these things come to the clinic and what they will bring will be whole new entire genome analyses that generate the same kind of problems that we're dealing with whole genome sequencing but in what, like string theory, 11 dimensions or something? So just multiply with the problems we have with the informatics and everything else by at least an order of magnitude and maybe more. Anyway, so in the human genetics we bill about $11 million and collect about seven and about the same ratio we have about a 60% return. We're sending out and this is a new assignment for me so I'm not really certain exactly what tests are being sent away but our pathology department is sending out about $2 million worth of work that is deemed to be genomic in origin and so at this point I'll talk a little bit about our pharmacogenomics testing which has been something that is clinically used to a great extent in our facility, largely by the psychiatrists and because of the difficulty they have and so there's some special issues about how that's happening and weaknesses and strengths of that program that we plan to work on. So we have 36 drugs that are actively being used and commonly used that we look at the Cib 2D6 and Cib 2C19 testing and then bring this back to the patient record with the test performed, the genotype, the predicted phenotype, whether the patient's predicted to be a poor metabolizer, an intermediate metabolizer, an excellent or an ultra metabolizer. And then dosing recommendations and as you can see on these graphs, the dosing recommendations change based upon what the genotype is, and MIPR means at one end which is extraordinarily different and these others are down here where it doesn't make as much difference. The testing limitation, sources of supplemental information and then an invitation to get a genetic consult from the pharmacogenomics service. So this is the website for the genetic pharmacology service and of course when they talk to our clinicians they're interested in getting the consults so this shows up. When you order one of these drugs, a panel appears in the electronic medical record to invite the pharmacogenetic testing and whether or not a consult is desired and so those things are automatic. It's done by drug name. There's about a two business day turnaround which turns out to be really important and so there's a large psychiatric admissions service. Maybe I think it's about 15% of our admissions or psychiatric, something like that. And the physician makes a decision about the initial dose for the drug and then the test comes back and altered or not according to the physician. We've been trying to fund an outcomes research project on this but have so far not succeeded. And then the consultation and other educational materials are available. We've done this on more than 8,000 patients and have done 1,700 this past year and have quite a large database, mainly in the psychiatry services and some of the autistic patients are tested. Physicians have different desires for getting this test done. Some of them want to understand the genetics and have a sophisticated idea about what's happening. Others just want a drug recommendation and some want the panel to cover what the potential drugs might be. So there's not much data to show that this is really a dramatically useful thing to do but it's now embedded in our practice in a way that our physicians and patients seem to be happy with. One of the measures is behavioral intervention score and we can, in these 300 patients with various things, mood disorders being the most common and anxiety disorders being next. We can correlate the phenotype with the number of behavioral intervention score changes, problems that the patient has in that initial period. And so these are the patients on drugs that are dependent on Sib 2D6 and Sibs 2C19. And the next slide shows another group of patients that are not using these kinds of drugs that are dependent on this particular route of metabolism. And so we have some evidence that it makes a difference. We'd like to see outcome data but we don't have that. We are doing the neuropsychiatric drugs, some of the codeine and related drugs, azathioprene and warfarin. And we have the capacity to do the drugs that are not pediatric but are very popular among our adult hospital colleagues. And then we're going to make an effort to bring the immunosuppressive drugs, tacrolimus in particular, and the mycophenylates along with the other morphine-like drugs and variconazole into the picture in the coming year. And we'd just like to emphasize that when we're, here's an example of using tacrolimus that we have three different cell types represented here. And each of these cell types has a different spectrum of molecules that influence the consequence of having the drug in that tissue. And so it makes it fairly complicated. We hope to bring to the neuropsychiatric panel an additional set of genetic testing that will be made available to the clinic in the coming year and expanding this neuropsychiatric panel. Whoops. And then we have decision support tools that help make this, integrate this into the medical record which makes it a little bit easier to deal with. Organizing our complex patient in an efficient way, risk stratification, and automatically lead to suggestions that would be helpful for the clinician. So this is for a kidney transplant patient with immunosuppression, cardiovascular disease, behavioral management, and chronic kidney disease, all being summarized in this form from the electronic medical record. And then risk current therapy, suggested actions, and the provider response, also on the same form with relatively easy to understand color coding that identifies places that are in trouble in red, things that are okay but cautionary and yellow and green for things that are fine. And so we are implementing, there's data outside of the EMR, there's still some manual input, and we are working on fixing the adherence. And so the pharmacokinetic data, the adherence, the protocol recommended drug levels, and the passively reported patient outcomes are now being put in the same place and hopefully incorporated into this and within the year we hope that this additional information will be available. And so we'd have adherence, rejection, survival, and cost all summarized in one place for quick and easy action. And so including social networking and smartphone apps that we hope to bring to the table and then apply this same strategic approach to other conditions. So that's basically our genetic pharmacology service. I'll turn my attention to our clinical services. We do these tests at the moment and they're examples of how they're useful for a whole variety of different conditions. There are 60 human sequencing genetic tests offered at the moment that use basically Sanger sequencing. We are converting these to next generation sequencing and these are the first two that are being used using a targeted sequencing approach. And so the strategy is to take the testing that we're now doing by Sanger sequencing, do that with next generation targeted sequencing. We confirm all results that we report by Sanger sequencing and then take the, there are 21 tests that are in the pipeline for being brought into targeted sequencing and they're listed on this slide and the next there's some three or 400 genes that are involved in all of these. So we expect that whole exome sequencing or whole genome sequencing for that matter will become inexpensive enough that it'll replace the targeted sequencing and that we will be moving eventually to whole exome sequencing on just the basis of cost. We expect that the next generation sequencing will finally become reliable enough that confirmation in the way that we're doing it now will not be necessary at some future time. When we do do the whole genome, it opens up the possibility that we can query the rest of the genome for especially for those patients in which things are not, where we don't find the answer with the targeted sequencing even when the clinician asks for a particular kind of syndrome. And then of course, we have to cope with the incidental results and we will wave our arms and say we'll follow the best practices at that time. My personal orientation is that we should be returning as many results as we can but that is not shared. We have the experience with the patients with Huntington's disease that initially really would like 80% or 85 or 90% of them say they'd like to have the results returned. They go home, talk to their relatives and they come back and it drops to something like 15% that want to know the results. So when you're bringing a new technology into the clinic, there's that issue about discovery versus clinical utility. And so as the children's tertiary, quaternary referral center, we see a lot of unusual pediatric conditions. And so we built the pipeline to do whole exome sequencing and at this point, and we're doing it on a semi-research basis. We have a clear approved pipeline but we haven't set up the financing yet to do other things. So it's driven, as Mark mentioned the other day, by the interest of the faculty. And so we've done about 366 subjects and we do trios in order to have reliable data and know what the de novo mutations are. I'll mention this condition. This is Barrettster-Witter syndrome. 18 of 18 children known with this have de novo mutations in one of two genes. And they have this intellectual disability, hearing loss, seizure, short stature, microcephaly. And this technology is going to be available for all kinds of rare problems. Since Mark is holding up his hand, I'll just mention that what we're hoping to do is to develop this pipeline and use it an iterative improvement way that incorporates genetic counseling, technical capacity, informatics, the interpretation and the financing in ways that will make this work as a clinical test for us going forward. We would probably begin with intellectual disability in children. That is the condition for which the literature supports the greatest confidence that you'll find a de novo mutation and autism. And then we have many, many rare and uncharacterized conditions. So we're in the past, we have Sanger sequencing and we're headed towards whole genome sequencing and other genomics. I expect that epigenetics, expression, metabolomics, chromatin immunoprecipitation, chromatin confirmation experiments will all be important for clinical testing at some future time. I would just mention, before one of the speakers said that, well, we're becoming experts in particular fields, so you're not a medical geneticist anymore, you're a cardiologist who's an expert in the genetics of cardiology. And I think my Howard here next to me was saying there was nobody in rheumatology at the meeting and I was squirming in my seat because that's what I do. And so, yeah, okay, okay. And so, this is a paper about the epigenome-wide association data, looking at using that 450 array from Illumina and relating rheumatoid arthritis to the SNPs, to the single nucleotide polymorphisms that have been related to its risk. And they found eight or nine examples in a small subset using this approach called the causal interference test, the causal inference test, excuse me. And putting differentially methylated sites in the middle, they find that they can go from the genetic origin through methylation that's different in cases and controls in a way that infers causation to the disease process suggesting that the low odds ratios that we have for GWASs can be improved by looking at these epigenetic risk and combination. And so, luckily for me, Terry is still up to two fingers and so I'm at my last slide. Thank you, John, that was great. Just as the apologist for the group, while I think we all recognize all of the exciting things that are coming, our orientation has always been very practical, which is what is here that we can actually do something with. So, I imagine we'll eventually get all of those things as people that are involved in the foundational work to say where is the utility, we'll eventually be taking those on as well. That's a certain amount of job security. Yeah, I think you'll have meetings to organize in the next couple of decades for us. So, a couple of questions. One is, I wanted to go back to that kidney transplantation example that you used. Two questions about that. One is, are all the data that are collected as part of that form then represented as structured data in the data warehouse? And then the second question related to that, as you mentioned, patient outcomes. What tools are you using to collect those patient reported outcomes? Well, I'm not certain. We were using all kinds of different measures of their hospitalization time, the time in therapy, the various tests for how their psychiatric condition was progressing. So, we were interested in developing the preliminary data that would make a real test possible. So, that's where we're stuck with that. The psychiatric practices vary greatly from practitioner to practitioner. Drug selection as well as how they interact with patients and how the progress. Yeah, I think it's really important because you mentioned specifically and it was nice to see passive patient outcomes and then patient reported outcomes. And I think we sometimes forget that those things are different, that how a doctor thinks a patient is doing and how the patient thinks the patient is doing are frequently at odds. And clearly, the movement is towards patient reported outcomes. And so, I think as we think about some of the implementation strategies, we wanna be able to move to collect patient reported data to get, to address that specific question and use where available standardized tools or where not available, think about development of standardized tools that can be used broadly. I think the, I just mentioned, I think one of the reasons that the psychiatric pharmacogenomics is used so extensively is because the psychiatrists have now a whole slew of drugs to choose from and they really don't know, have any idea which drug to pick. And so, they're kind of flopping around a bit to figure out which ones to start with and how to make that work and what little bit of help they can get. So, it may not be that pharmacogenomics adds so much, it's just that they have so little to discriminate between the various choices that they have. Yeah, this gets exactly to what Ned is saying, that identifying clinical context where, again, you maybe have therapeutic equipoise or you have a number of things to select from. The barrier to use additional information like pharmacogenomics to inform your choice and then measure the outcomes doesn't really add much in the way of risk to that. But it may be very important in terms of the outcomes and so, it's a great, I've been really impressed way back to 2004 when I first started hearing about the work about the opportunity that we have in an area that really is not being explored particularly well. And in fact, EGAP has gone on record as saying, there is in fact no reason to do CIP-2C-19 testing whereas I think you have data that I've seen presented at least informally that would suggest that that probably is not the correct conclusion and we may have to revisit that. So we're looking forward to more. So, Jeff. Yeah, I wanna congratulate you also on the pharmacogenomics consultation model for neuropsychiatric medications and the fact that it's been around, I don't know how many years marks at 2004 so quite a long time suggest that you've figured out a sustainable economic model for that to work and could you comment on that model? I think our reimbursement rate for that is about the same as it is overall. About 60%. So we think we're doing pretty well with it. How is that, is that being billed then as a separate professional service through like a pharmacy type of a... I'm not certain, but I think it's billed as the test is billed separately from the consultation service. And so the test and its automatic report comes back and that's billed separately and if you want a full consultation then that comes from the Pharmacog Genomics Testing Service. Okay, great. So we will take a break. We're going to, oh, I'm sorry, Maureen, I'll... John, I just wondered along the same lines if you looked at whether there was any difference for the physician in terms of length of time they spent with the patient? No, I mean whether these are... Whether that extended their time in terms of having to explain things or talk to the patient with having the Pharmacog Genetic Testing or they felt that maybe it was more efficient? That kind of randomized control trial which we would like to do hasn't been done. Okay, thanks. So we will take a break. Again, I think to be sensitive to people's time we will do it for 20 minutes meaning we'll reconvene about 10-2. Thank you. Thank you.