 So maybe I'll start the discussion with just listening to the two, the give and take. It struck me that a lot of what you're doing in your lab, I mean normally in clinical medicine if one is sending out for a test and I'm going to be simplistic and sort of bend this into sort of esoteric testing, you send out a test and the lab doesn't machinate over this the way you do, they send it back and then you consult an expert, another clinician, hopefully a local clinician expert or you may even refer the patient to that expert to sort of go through this and it seems like there's a lot that is kind of being dumped on you in terms of going and looking up the primary data and figuring out why some other lab had a different interpretation and so I guess it throws out a question of is there a different approach that maybe Caesar could catalyze research on around different ways to handle the dissemination and in sort of administration of this kind of test that doesn't sort of rely so much on the labs because yes over time we obviously need to have databases that are accurate but we all recognize that this is still a rapidly evolving field and so maybe there's more of a balance for consultation that's not so heavily in the central, I'm just gonna throw that out there's one thought but maybe that's maybe there are other other questions as well. Heidi. So I think that you know one of the things particularly as we've all gotten it into exome sequencing and genome sequencing you simply can't interpret an exome or genome without clinical phenotype I think one of the challenges that many of the laboratories have had has been the notion that you just send the DNA and out will come an answer and you know without a phenotype that can be challenging and in fact I think the phenotype can help build and inform the interpretation so I think thinking about better ways to have that be a you know a dynamic process you know and I often hear of physicians who get clinical lab reports and you know I think Gail made a comment get a lab report back I disagree with it but if they don't go back and tell the lab why they disagree and provide that information there's not that ability to iterate over time and so I think it is and I have seen that happen within Caesar I think it's a great environment that Caesar has where there's a dynamic between the laboratory and the clinicians that's informing this over time and I think that's been a great example and and I would actually argue getting back to the you know original bake-off project that we've been doing although we are seeing differences in the implementation of the guidelines across groups having watched Caesar evolve over the past four years the groups have really developed and improved and learned how to interpret variants in a more sophisticated way than the start of the project and I would argue that our ACMG guidelines right now are a bit dealing with another area of laboratories that are just still reporting variants as pathogenic because somebody published them as such and and I think we've managed to get sort of that really poor environment addressed and I think that guidelines help with that and now we're sort of looking at the implementation in a more expert environment which I think Caesar represents and now we're in a next phase of okay how do we move another tear up and improve the quantitation and Gail pointed out that you know some of the things that we're starting to think about to make these guidance guidelines more objective and less subjective I add to that that it has been very simple for a lot of these rules to come to a general agreement about how they should be used even though we came in maybe using them differently but to agree on a framework of what was what was the intention of the rule sometimes the literal wording of the rule is a little bit misleading and so we have been able to reach consensus about what these rules means I think feeding back to the community then about what is the consensus interpretation of the rule is going to be a really important step forward and Heidi you can bet when I don't agree with someone's lab report they hear from me let's go over to the other side of the room Pilar thank you so I had a question for either of the speakers or anybody else in the room do you see differences with respect to sort of the field that are ordering tests so the clinical domain in which the test is being ordered do you see differences in different clinical domains with how the ACMG recommendations are being interpreted or how sort of at home rules for calling variants are operating and I asked this question because I I'm not a clinician but I consult with some clinical labs and some clinician researchers and I feel like I see differences in the amount of evidence that different fields of medicine require before they would call something a pathogenic variant or not and I'm wondering if if the medical practice in genetics is creeping into the laboratory practice yes potentially and in in doing these ACMG and P guidelines our goal was to try to do it across the board for any gene and you know acknowledging that we need to get into more disease specific I have we do get a lot of genetic testing from non-specialists the ones that seem to to work better with what genetic information is neurologists seem to know it extremely well obviously clinical geneticists and when you get into more primary care physicians as when they they are going to rely more on the laboratory than what some of the specialists will be doing but I do agree that it really is going to depend on the condition and what the other genes are and a whole bunch of different parameters of how strongly I will feel about calling something pathogenic versus likely and I also I know talking with our clinical geneticists I was surprised that for some conditions hey you know if you can give me 30 to one odds I'm going to take it another one was saying no I've got to have a hundred to one odds before I do so if there's a lot of variability there yeah I mean I think that it's been a novel thing in medical genetics in the last five years that we are reinterpreting the variants at the clinically you know in ten years ago if they said it was a pathogenic BRCA mutation I took their word for it and we're in a different phase now where so much less is known about the very rare variants we're finding on these exome and genome tests there's very little evidence base and there is such variability among what the labs call that you have to actually look at the primary data and we've had to retain even not just our medical geneticists but our genetic counselors you know now can get on XAC and look up the allele frequency and we one of the things we also feedback to labs is when they don't give us enough information in the report so there's certainly one lab that just says this is what we're calling it we're not telling you why and we try and avoid that lab for you know because we want to know why they called it that way to see if we agree with that reasoning it's the same thing as for what goes into ClinVar you want to not just see what they called it but why they called it but realistically medical geneticists loses money because we spend like an hour with every patient we spend an hour in prep time we can't really spend a half hour of prep time looking up all the evidence for the variant and sometimes it's four or five hours to really go through the literature for every patient that really needs to go to the lab and we need to be confident about the results that we're getting but we're not there yet but we can't continue as a practice without fixing that you know so some part of this I think is are you know we ready for for prime time and I don't know when I look at the results of the bake-off I don't know whether to laugh or cry obviously we would all want it to be what want those results to be better but then another part of me thinks are we holding ourselves as geneticists to a higher standard than the rest of clinical medicine because I also practice in cardiology and we do echoes and that's been very established for 30 years and I regularly like every day in clinic gets and echo reports that I don't agree with and so I also wonder really it's a question of timing at what point are we are we ready to unleash this and just be comfortable with the uncertainty yes so I actually do the QC for the carotid ultrasounds and there are people disagreed by one classification commonly right and that's like considered just fine and there's only four categories for that so I think that we that it is important to say we're not going to match all the time maybe the goal isn't to match all the time but the goal certainly is to have a set of useful criteria that we that we all apply in the same way and I think that's where we're trying to get not to total agreement but to use the right criteria in the same way or we had a question back in the corner here yeah Bob no spam I actually had three comments to make I think what one problem is that talking about a one step doesn't really help that much because the step between pathogenic and likely pathogenic is not the same as the step between likely pathogenic in V us whereas you have great gradations of your carotid and degree it's just a different it's just a different quality and I think overcoming that is really one of the major major challenges the second comment had to do with about that leave are you raised about consulting with the laboratory the vast majority of genetic testing these days is not coming from academic centers it's coming from for example in the cancer world particularly it's coming from private oncology and even more so private gyn and oncologists that's where the vast majority of market share for the company that shall not be named is coming from and so they those practitioners have to have some expert to call and talk to and I can't think of any anyone who could do a better job of interpreting their result to the patient than the lab to the practitioner the lab that generated the result so I don't think we're gonna get away from that until the level of education of non-specialists rises and then the last point I wanted to make had to do with are there different practices among different practitioners and I think it's inescapable because in many ways the interpretation of a variant although this is certain to some extent a circular is also highly Bayesian so if you have a patient who has a dig dig a an obvious cns abnormality and a more tooth sign and has you bear syndrome and you have a typical gene that's involved with a variant that maybe it's not quite clear if it's a if it's a variant of uncertain significance or likely pathogenic that's going to be looked at differently than when you have a disorder where there's a very high prior probability that it's actually not an hereditary disorder at all sure well I guess we go back to Eric's comment earlier I think it's been a fascinating discussion but maybe in the wrap up people could say well what is Caesar 2.0 gonna do about this problem is it the question of improving the guidelines or better consistency or facilitating ways for clinicians to communicate back to labs I agree with you they think the Caesar sites all communicate with their labs actively but like how are we gonna help solve this problem so Sharon you're setting up Deborah's closing remark which is great but there was one more question in the back and then we'll go to the closing remarks Naomi Aaron some blue cross and blue shield association I'm very impressed with the quote one gene at a time one variant at a time with careful investigation but there's also a contrast to what's being seen commercially as health planes were encountering this frequently and that is very heavy marketing of profiling and paneling in certain clinical areas but not confined just to oncology very broadly and it's actually the existing of sequencing itself that makes this feasible to offer these are going directly to the clinician perhaps are even more attractive because they come with the interpretation all rolled up and my question is how will a CSER approach this question beyond one gene at a time in sequencing to what is the relationship the algorithms the proposed algorithms there is something very concerning here too because there can be different panels of tests different profiles that come out with sort of the same predictive capability but not necessarily selecting the same patients you can think of a positive explanation for this maybe the target is broad enough to accommodate this or a negative explanation maybe there's just a lot of air here but it does I think force us to start thinking about the interactions I just wanted to raise that question I would direct you to Carlos Gallagos paper from CSER which is on what is the optimal size cancer panel what kinds of genes what number of genes it was a really nice a work that David Bainstra here is the senior author on and that kind of outcomes work I think we're gonna have an outcome session later is critically important and a very important space for CSER I would agree all right well that was a great discussion so now I think it's time for the closing remarks by