Create Mark Data with airSlate SignNow

Create mark data

um so I'm mark from Sala bio easily impressed by technology and so first of all thanks so much for having me this is it's a really great event there's all a lot of people and really combines all my interests health data people listening to me talk so this is this is really awesome and the other the other presenters have just been amazing so hopefully hopefully I won't be boring so first up a few words about what we do to give some context so we want to be the data infrastructure behind genomics so what that means is we want to use data to help interpret the genome for clinical and health purposes so if you go to the doctor and get or good we combine and get your DNA sequenced it usually involves ingesting a lot of data from the outside world about what different DNA sequences mean and we help provide that data we put it we normalize up in the right format quality control etc so and we think that that's really important since genomics is growing quickly so I want to talk to that today a bit of a perspective on kind of a new a new buzzword that's going around I called precision medicine and where I think genomics fits in and where we're data scientists fit in and and how I think together that those communities can really make a big difference precision medicine is the idea behind it is that medicine is becoming more and more data-driven it's no longer supposed to be about guessing more about you have a bunch of data so now there's evidence that can be used cetera so people are investing a lot in developing technologies to measure a lot of information about people the problem is the the big concern is what if precision medicine actually just becomes we have a lot of data but we don't know what to do with it yet medicine so why measure all this data why sequenced your genome and literally why sequence your genome I mean for most people there's no information that they gleam yet from it and you know they're the big fear is that kind of looking back 10 years from now that miss precision medicine look a bit more like like the past than the future that's actually my doctor so don't laugh um and you know the question is what exactly can we do with all this data that's being collected how does it how does it actually translate to something that's you know medically actionable and right now we're not exactly quite there yet so there are some definitely bad news and headlines first missteps about using genetics a lot of people are scared about it so there's a lot of really promising new technology Alex gave a really great background on on the field of genetics this is a an aluminum ishino extremely really fancy this thing can sequence your genome for $1,000 and ten years ago cost like three billion and was really hard now you just get one of these machines kind of looks like an apple I think it looks really awesome and people are buying lots of them Columbia has a bunch of them Harvard other prestigious names have them lumen is doing really well there were three billion dollars they make these sequencing machines because people are just super excited about using genomics for stuff and that's really cool so you would think that and also okay there's also this thing wearables people are this can tell you if you have a pulse or not how many steps you've taken and whether you're sleeping so um so and then there's this thing just launched looks really cool I don't know what exactly does yet but this is it's really awesome also looks like Apple I think I think we're getting somewhere um so this is awesome so you it you would think that this is what the doctor's office looks like this is your doctor now it's what I when I sings a Bloomberg terminal and there's all this data coming in real-time I know everything about you all this your DNA your sleep schedule all this whatever Fitness stuff that you do and your doctor's there he's eating lunch and he is deciding what to do with you that's really awesome and that's oh my god this is actually what it looks like right now and I this this is literally what the data behind genomics behind precision medicine looks like right now FTP sites I know everyone just got sad looking at it this is this is my life this is a lot of genetic so every genetics company every hospital they design and implement tests to figure out what disease you have or don't have or whether you should have a child or what drug to take based on your DNA they use this data that's collected by academics and deposited in folder called data somewhere on FTP sites and this looks like the 80s but this is actually from like last week that's that's true this is true so 23andme uses this data and papers council Rika Mize I don't speak for you but you know everyone uses this data and it's it's valuable data this is this is data that academics have collected over the years and its really great and there are even instructions for how to use it it's it's a readme file so it's great data for the first kind of chapter one of genetics but the problem is that there's no quality control you don't know what which records are right which records aren't relevant there's no feedback for this work for me this didn't work for me there's no system of knowing which version you used which you didn't use and this if you if people in the industry are worried about a regulatory backlash or consumers being underserved this is definitely one of the pain points and scary and frankly yeah it's really scary there are some great companies in the space that are doing a great job in the software ends as Alex mentioned but then there are other companies that might not be data scientists they might not have software I was really scared to go to those hospitals or those companies and know how they're using the data from the outside world to determine what disease you have or what you should do and where's the time thing I know somewhere should be looking somewhere okay just let me know if okay so that okay so the main kind of thing that we're interested in solved by oh and I think as a community were we're interested in is we've been able to generate as a community a lot of value from knowing a very small amount of information what DNA variants the things that make us all different which ones cause common diseases and there's still a long way to go to being good at measuring those and there are there will be multi-billion dollar companies that will figure out how to measure those efficiently deliver that information to the patient and get feedback from the patient there's a lot to do there but the interesting thing is that that's just the tip of the iceberg tip of the tip of the iceberg it's it's crazy there's so much information that we don't know what it means for your health or for medicine and if we can collect and understand that information so for next few years were just the genetics world because there's so much capacity they're collecting so much information they're observing so many new DNA variants and none of them are being understood or explained or very few of them are that's a huge opportunity if people can understand that then they could truly transform health and I want to tell a quick story equal example about a way that I think I think this could work a way that I think is kind of a the illustrates how we can find out this you know the kind of known unknowns okay let's do it so there's this this guy in Brooklyn that we know and he is the head of this rare disease foundation and families that have just crazy inherited rare diseases that hospitals and doctors can't explain go to him to try to explain the disease using genetics and the the set up is that there there's this family and he told us a story but this happens to a matter of a few weeks it this family goes to prestigious Hospital America has their DNA sequenced and they come back with yeah we don't know this could be a fatal disease we don't really know we have these two DNA variants that come back that make this person unique and there's a really really bad phenotype but we don't we can't explain why these DNA variants cause this phenotype because they haven't been seen before right so the variance are unclassified that's what they're call the diagnosis is unknown the family doesn't know what's going on and but the hospital has done its job right and the normal process there is that it takes years of research to understand why this particular DNA variant causes the disease that's just how that's how we've been as a society collecting information collecting knowledge and so so this rare disease foundation was a year into that process when another family with that same exact phenotype went to a different hospital also they came back okay unknown unknown variants and that family went to him and he was able to just match up on his computer these two datasets and he saw that the variants were exactly the same and so because were two totally separate cases that that allowed him to explain this disease that this variant caused this disease so instead of years of lab research building a molecular model and zebrafish in my sort of he was able to use two separate pieces of evidence together now they explain this disease that's this is this is huge because all of a sudden just by aggregating two datasets two very small datasets examples with standardized evidence you're able to make a new discovery that explains the disease people can then screen for it later like create drugs for it and this happens a lot and the take-home message is that if people can aggregate genetic and clinical genetic data with clinical evidence in a way that people can use other people can use then that will completely change the field and really explain a lot a lot more and so the government the government is engaged in something like this and we're supporting sort of of it but it's just like another FTP site where people submit things and the problem with that is that it doesn't work with everyone's understanding of privacy there because it's just a public site it doesn't work with everyone's understanding of IP because some companies collect data they consider their IP some companies don't different attitudes there and also it's just an FTP site so people dump data in there there's no real incentive necessarily to do so some companies do it for branding sometimes publicly funded research they do it because they feel like they should I buy no one trusts other people's submissions we found because the standards are different it's just an FTP site and at least the process for submitting is easy you just have to fill out 250 columns on a spreadsheet so that that part is fine that part that part they've solved but this this is a this is a huge opportunity for people that know how to standardize datasets parse munge data and we think that we think that kind of the things the data infrastructure for genomics of the platform that we've built can be helpful the idea is that there a lot goes into take going from a DNA to a clinical report that's software alex has mentioned a huge part of it there's also the ingesting data from the outside world which is where we where we help and we think that if there's a way to build in a system for people exchanging information distributing it in a way that respects their privacy and IP and in a way we're too siloed institutions or individuals can come together to form something that's together valuable wear separately isn't particularly valuable that that can really drive the field forward and you know as a contrast to the FTP site first of all you we're opening up our beta you can sign up interestingly I don't think you can access it from from within Bloomberg all right I think it's a conspiracy that's the only explanation but in contrast to the FTP site we make it a little bit easier you can access public data like the the clinical data that I showed you with just one line of code because we made an API to all that data we have really great interfaces so they could explore any DNA barrier you want with all the data that's available to annotate it from the outside world so anyway we're better than FTP site what I'm trying to say okay so that's that's where we are email me mark it's all Viacom follow me on Twitter or in real life if you're confident okay that's okay thanks thank you Mark so the fundamental driver why people in the clinical world and and other parts of the ecosystem would want to work with a platform is that it's not really better better tools better experience because it's been very hard historically to make all these people work together it's not just a question of yeah right exactly so there are there are there are a number of factors so when it comes to ingesting data yeah there there are there is one approach where everything can be built vertically so you take all the data that's available from the outside world yourself and maybe you make partnerships with people that provide that data and up till now that's in some sense been the case 23andme has their own curation etc etc but the way we kind of see it is that that that inflection point has already happened there's somewhat akin to the microprocessor that forced a lot of horizontal ization the aggregation administrating that's happening in genetics and this is a great building to be in to talk about that because that in finance that exact thing happens I mean the idea of a platform delivering data to an industry and making every player an industry better I think that's not not too crazy when you're sitting here so that's that's one of the motivations that great as we getting close on time I want to open to questions so so what's what's on your what's on your roadmap I guess what are you well the main thing is figuring out all the right all the right incentives and the the security structure to be able to allow people to expose data to partners but kind of have access controls so that people that they they don't want I'm seeing the data won't see it so that's that's I think that would actually think of a really long time that's kind of a work well we'll be working out for a long time great yes crowd that really wants to know about money yeah I guess I'm just trying to understand what the revenue model is yeah I feel like that's us Ron that also um wait yeah so what is it like 2006 okay I so we think that there's a lot of information that and we don't think we know there's information that people want to monetize and they need channels to do so so people collect data they create their own interpretations of data and they want to distribute it to people that use that data so farmer wants to consume data that someone else has collected we've seen a 23andme did an enormous deal with from data that they that they had the reason was valuable because they had all the proper consents for this fairly rare collection of variants and clinical data associated with it that's going to happen more and more and series starting to happen and there are different kinds of data there's clinical evidence and as much information that you have there the sort of the more valuable that is and that will never go away there's always more information that you need and and then they're just kind of interpretations on top of the genome and also those are things that you want to kind of distribute it's is similar in some sense to Bloomberg there's a platform and a channel to distribute information and if you have a good channel to distribute information like Netflix or Bloomberg etc some of that contact can be open and sometimes premium soon as you make your own so I think think things like that great one more office or tea okay yes so I know that from your presentation you really made a point of the quality of the data and how the data is accessible and knowing with clinvar and Omen and CBI and all these sets how are you guys really controlling that quality how are you doing the QC for that data yeah so the I mean the first thing that we do the first part of that the accessibility and just the kind of the first level quality level one is we kind of reindex the data and normalizes so as you know everything is kind of in different formats different sees different versions it could be this build of the genome different builders you know missing certain fields has extra fields for some reason or just like the whole records missing that so we fill all of that in so initially it's been a somewhat manual process on our ends it's taken quite a bit of work but the sort of next the kind of layer of that is that interface that I put up where you could actually explore variants it has feedback mechanisms and so it I think one thing that I kinda knew that I like to make is you know if the human genome was quote-unquote mapped in 2000 I think there's kind of one still needs to make the google maps of that and then one after that we would need to make the ways of the humans you know where you actually have feedback where you know something is correct not correct from patient in real-time way and I think that that could be really promising okay terrific works thank you all right thanks a lot