Awesome. Well, we're glad to have you here. And I understand. So, where are you coming to us from this week? >> Uh, Sydney, Australia, >> and you made the trek all the way here from>> Very happy to be here. First tech conference.>> Oh, wow. >> First tech party last night, so a little bit flat. >> Yeah. So, how does that compare with the healthcare? Uh, healthcare? Uh, healthcare? Uh, >> well, IT conferences. >> The genetics conferences, a lot more ladies. Dance floor is wild, but the dance floor was pretty good last night still. Yeah, [laughter] >> the genetics one. That's the >> genetics conference. You got to get to one of those. Yeah, that's a lot [laughter] of fun. >> That's the best kept secret. Uh probably >> 100%. That's Yeah, take that take-home message.>> Awesome. So, um speaking of which, uh I know we're a little bit thin on the show floor here, but for people that are in earshot as they're walking by. Um can you give us kind of like an elevator pitch? Like what's the value that they might get from listening to the presentation here? >> Yeah. Well, today I'd like to tell you about a lot about precision medicine, how we're using genomics in real time to guide care of patients, get them the right treatment at the right time, show you some little sneak peeks of the impact that has on a couple of the children that we've seen. And we've now applied our process to about a thousand children, and we'll show you the systematic improvement in health outcomes from using data to try and cure this disease. >> That's amazing. And when you say using uh genomics in real time, that sounds pretty data intensive. Uh >> yeah, absolutely. a lot of cloud computing, quite large, quite sensitive data,>> challenges in how you share that data with researchers because it's publicly funded research. We're obliged to share it.>> So, extra wrinkle. Yeah. >> But then it's obviously pretty sensitive. So, your genomic code is not just your genetics, it's your family's genetics. Andso that's a difficult thing.>> Yeah, that's uh that's a great point. I didn't even think about that. All right. Well, come on in if you're passing by. I'll take another minute or so to see if we can talk to anyone uh into coming by. It is a little quiet here. So, but we're glad you guys are here. It's going to be well worth it. And we're so glad you're here to finish things out for us here in our final presentation. And um yeah, if you're passing by, I see a couple folks there. You'rejust in time. It is our final presentation of the week here in the IT Factor stage. So, if you haven't caught one, you want to make it happen right now. And we got a great one for you. We're going to be uh talking about using data in the fight against uh children's cancer, which uh is about as important a cause as I can think of. For sure.>> It's a feel-good story for sure. [laughter]>> Yeah. >> Yeah. >> Yeah. >> Right. Um all right. Well, that didn't seem to work. So, rather than me just up here uh flailing away. I know you got a lot to share with us, so why don't I just get out of the way? And again, thank you so much for being here. Thanks, everybody. Thanks, everybody. Thanks, everybody. >> Thank you, Paul. Thanks everyone who's come. This is nice and intimate. Um thank you. Don't worry. All right. Yeah. What I'd love to tell you about today is really, have any of you really experienced genomics data? Understand what cracking the genome is? I thought so.I'm going to try and start there. What's the genome? Why do we care? What's the value in understanding it? tell you a little bit about our national precision medicine program called zero childhood cancer and how we're using genomics toimprove health outcomes or some of the go deeper dive into the data ecosystem that we've built to try and facilitate this andI'll touch on some successes as we go through. So all right so this is my team. Everything in science is built on teams. We start with the people. I just wanted to highlight though the you know I'm a researcher a bioinformatician. It's kind of like a data scientist that understands biology but my group has just diversified over the years. So now we have software engineering, data engineering, data science, bioinformatics and our most recent hire in UX development. And I think that tells you a little bit about the nature of how my industry is changing. um from building tools to solve specific problems to having a really holistic approach, lots of partners which we can't do any of this without our partners. Um so what do we start at the beginning? The fastfacts about childhood cancer is that 60 years ago this was a death sentence and in fact the very first cancer cures were in children with cancer driven by a lot of research in the US. It's the leading cause of death in children and young people in Australia, in America and other developed countries. There's about 400,000 children globally affected by this disease, about 16,000 in the US. So, it's a pretty big health problem and we want to do something about it. It's now in some cancer types like leukemia, 90 plus% of them are cured, but those patients are cured by really toxic chemotherapy, radiotherapy, lifelong health impacts from those severe treatments. So even though we'vemoved the needle on cure, we start to think about safer and less toxic treatments so that thesepatients uh these survivors actually get to thrive. Um so yeah, that's our challenge is to simultaneously improve the outcomes for the children with no treatment options and then have safer treatments for those thatdon't. So don't. So right. So yeah theytypically >> yeah that'sa good question. So kids are not little adults. So their cancers are different. Many of the adult cancers are from exposure. So ourlifestyle last night UV damage. But for kids they're developing their organs. So there's these trajectories of the cells that have to divide precisely. they have to stop. So, a lot of it is the breaking through of developmental checkpoints. And so, as their little organs are growing, they just suddenly they they've missed the cue to stop. Um, and then they just start growing out of control. So, yes, adults have acute lymphoplastic leukemia and so do kids, but they're fundamentally fairly different diseases, whichmeans it's hard to treat them. Yeah, we need a different approach to treat them, which is pretty good segue here. You know, when I started my career in this space about 20 years ago, we were treating all breast cancers the same, all prostate cancers the same. But now, when we understand that a unique molecular basis of every patient's tumor, the genetics, especially the genetics, we start to get better clues and that allows us to match new treatments that target specific genetic changes to match each patient's unique tumor. Um, so that's been the big breakthrough in the last 20 years, the move away from the one-sizefits-all approach to something much more uh precise and matched to the patient. So, if you're wondering what is the genome, I've said this in a couple of forums. Um, it's a three billion letter code, and the alphabet's simple. It's just A's, C's, G's, and T's. There's three billion of them, and you inherit a copy of that from your mom and your dad. So, when you're sequencing a patient's genome, hello. um it's a six billion letter code that you're trying to work out the order of and when you look deeply in it when you print it out in 10point it'sliterally that large and so the challenge is to work out which are the spelling mistakes in every patient's genome and they help shape the way we look uh the way we respond to certain things um and our predisposition to different diseases. So, for decades, scientists have looked down microscopes on the left, and you can see the chromosomes in pairs, one from mom, one from dad. And in a normal cell, that's what it looks like. Nice pairs of chromosomes. And for decades, we've seen in cancer, these things get damaged, chopped in half, glued together, and these marker chromosomes at the bottom arederived chromosomes of bits of DNA all stuck together. So, it's chaos. But now with genome sequencing technology, we can see not just the high resolution stuff over here. We can zoom all the way in. Every one of these dots is a single A to T or C to G. So we can go from the single base resolution all the way to the big deletions, amplifications, little tiny bits of DNA. They're super sharply amplified.So the first human genome took 13 years and billions of dollars to solve. And now we can do this for a couple of thousand dollars in about three days. And so that's shifted the dial away from can we do this to in what setting should we do this and um and how to that's why the data scientists are so important is how you actually getall this information out of here. So in precision medicine when we're using this type of test for every patient we we're born with 5 million genetic changes. So every everyone in here when we compare to the reference genome there's at least 5 million changes. Um which zero one or two of those predisposed you to your cancer? That's a challenge we have to work out. And then in cancer, you pick up more mutations over time as those cells divide past those checkpoints that we talked about. The DNA damage checkpoints become kind of overruled. Um, so there's a thousand to a million more genetic changes and whichsmall number of those do the oncologists need to know about and the pathologist to help diagnose the patient. So that's what we try and solve for every single patient in real time. And I've shown you what one tumor looks like, but this is eight different tumors of the most deadly brain cancer called the diffuse intrinsic pontine gloma. These kids have like five month survival. You're lucky if you get to eight or 10 months. And for many years, we've known about these four key genetic changes. But you don't need to be a cancer expert to see how different all of these tumors look. These ones have lots of these marker chromosomes created. There's this local damage here. We call it like a contigious. Um they're very different. And maybe this is why uh why they don't all respond to the same therapies. So by understanding the unique jigsaw of every single patient, we think we can match thetherapies toeach patient properly. So that's a big take-h home message. Every patient's tumor is unique. We need to do the genomics and other tests to try and understand that. So that's a good segue to our zero childhood cancer program, which is Australia's national precision medicine program. We've now recruited over a thousand different patients. And until this year, we've been focused on the really high-risisk patients over here, the ones with less than 30% chance of survival. And we've now on the backof that success been able to launch our new study, which is for all children with cancer and try to see if this approach is equally effective for children who would otherwise survive their standard of care therapy. Remember I told you there toxic. So whatcan this information do to help? There's lots of science stuff here. It's very busy. We were very pleased to see our publication like in science you have to publish our findings but we were recognized as a top 10 advance in medicine by these big American luminaries. So wewere thrilled about that and in a nutshell when we apply this technology and over 90% of patients we find the genetic basis of their tumor. Um over 70% of patients receive a personalized diagnosis and a treatment plan. um and we doubled the rate of patients which had an inherited cause for their cancer. So 8% of patients we thought had a genetic reason in their family for developing the disease. Now it's actually 16% maybe even 18%. So it's really important to know this that tells you about the risk of cancer in the rest of the family. Um and importantly 60% of the patients who took thetargeted treatment recommendation had a response. Um I'm going to tell you what ex what that actually looks like with some case studies. Now um h so this is the process. A child gets diagnosed in the hospital uh diagnosed with cancer and then we get a little biopsy of their tumor or the whole tumor gets chopped out and we take uh take a piece of that down to Children's Cancer Institute apply all this genetic profiling technologies and some new blood tests that we're developing um and try and return that information in real time through a tumor board. There's about 300 clinicians on the mailing list which um and about 70 of them will come every single week to discuss each case which is amazing. Um so we've been able to influence the clinical care in that way. Um yeah and we try and change their treatment plan through that process. We also try and grow that tumor in a mouse or in cell line in a dish so we can then do drug screening and then we've got complimentary genomic based recommendations and drug screening based recommendations down here which takes a little bit longer. So let me tell you about Ellie. Um Ellie came to us when she was 11 months old and she had a tumor the size of her chest mass. Um theoncologists and pathologists couldn't diagnose her. They thought she might have had atumor called infantile fibro saroma but the standard pathology tests failed to find thiskey genetic change which is what they were looking for. Uh it wasn't present in her tumor. So she was undiagnosed treated with chemo uh basically in paliotative care at 11 months old. So what we were able to do was decode her genomes the fastest we were able to do it at the time. This is quite a few years ago now. So wetook the biopsy same day, got it sequenced that week, ran it through our analysis pipelines and found this new genetic change, the spec 1L. Oops. Um anyway, a particular genetic change that's similar to the one that they look for in the pathology lab, butnovel, brand new. Um so this helped us diagnose the patient and two weeks before she got diagnosed, a new pharmaceutical company developed a drug to specifically target this type ofgene. And so this is Lockxo Oncology now owned by um Bayer. And so she just had the most remarkable response by getting the firsttreatment in Australia to get this particular therapy. And five years 5 years later she's disease-free. Umit's not just this one case for Ellie though. We've got I think that's more than 20 patients now with this particular genetic change and we've got this amazing drug to combat those particular diseases. So I guess that's the other this and this was a different case where it was about a similar turnaround over Christmas. We issued the report on Christmas day which helped ultimately stop a surgery that was not longer no longer going to be needed and an amazing patient response. So that's the other take-home message is this approach of matching treatments to genetics. We need lots of different treatments to do this and I think in this next phase it might actually be combinations of treatments so thenumbers get a bit out of control. So, um, it's working. And then there's Jack. His genome looked pretty simple. Not that much going on in it. Um, but he had amutation that's common in melanoma. So, we ended up using a melanoma drug to treat his brain cancer and he was also in paliotative care andimmediately responded to this pill um, which had already been developed. So, we were able to repurpose drugs developed in breast cancer ormelanoma and use those in a different setting. So, that's a lot of what the genomics can unlock is we've got all these tools in the arsenal, all these different drugs, and the genetics helps match them to the patients. Um, I'm going a bit slow, but I know there's I I'll skip a few things and get to my main points. So, pivoting a little bit towards the data and how data ishelping us do this. You know, by the time we've done this in over a thousand patients, we've seen over 130 million genetic changes because there's just so many going on in these patients. We've done all this complicated genetic profiling, drug screening. The clinical data is probably the messiest part of our data. We try and capture it. We spend millions of dollars actually with paying nurses to capture clinical data and put it in our system. That's a huge bottleneck inwhat we do. And so there's a lot of data now. And so let me tell you a little bit more about what we're doing with that. So off to the data ecosystem. Oh, what have I done? There we go. Okay, so relooking at that process from diagnosis and relapse this time putting projecting which bits happen in the hospital and which bits happen in a research setting. We have very little control at all over what happens in the clinical setting which is very difficult for us. Um but as we start to capture patients, we capture the clinical data at enrollment. We do all the sample tracking and quality control data genomic analysis. We have to interpret it, put it through a molecular tumor board. We're generating lots and lots of data as we go through our process. We've been building a knowledge base and um and lots of different distinct data portals to try and present that data differently fordifferent researchers. Um so I'll show you a little bit more about that. The first thing we did about 5 years ago is adopt a hybrid cloud strategy. I think I've not heard of any other medical research institute in Australia at least that have done that. So we partnered with Equinex and that'sbeen an incredible success for us. And then we soon after partnered with uh withNetApp toget our storage grid and that's helped give us some data discipline because I had these large data files all over Australia and national HPC facilities, local servers all over the place. So we were able to consolidate that with our storage grid uh which has been fantastic for us and connect that to um to I'm getting ahead of myself. I'll tell you more about that in a sec. Um, so you will all be very comfortable with the idea of connecting your data to commercial cloud over here on the right. But what surprised me is that we could actually connect the storage grid directly into our national high performance compute facilities. And instead of staging up these analysis jobs in a very tiny scratch space where an engineer would have to put that file in the little two terabyte space and launch Q sub jobs on the HPC, we were just able to launch the jobs orchestrated by this new Nexflow uhNexflow workflow language for running these data pipelines and um and that was fantastic for us to be able to just queue up 100 jobs, 200 jobs, a thousand jobs and the data just moved across when it was needed. Um so that's really been a bit of a game changer. It sounds like a simple benefit, but it was a really big game changer that turned a 4 month horrible job into a week job. Um, so that's been going fantastically. Uh, I've tried to talk to a few people. Weuse this technology called Nexflow or common workflow language to chain these data processing pipelines together. I think this process has about 28 different steps in it. 5,000 real hours compressed into 24 by parallelizing things. So, we use the cloud a lot and we've been able to actually more than quarter the cost um sinceI started this. So, it's more like from $250 down to under about $40 since we tried since we started optimizing it. Um so, I won't go nerdy. A genomics audience likes this stuff. Um but yeah, that's how we process all that raw genomic data. So, I've given you a hint that as the patient progresses through our processes, the very heterogeneous data, there's clinical data. Uh we don't actually have any imaging data yet. Um butthat would be fantastic to start to get these drug screening responses. How did the which treatments did the patient receive? Did they respond to those treatments? So we're building a data lake this year. I don't have a huge amount to say except we're trying to see if this will solve our problem of being able to harmonize all this heterogeneous data. And then we have been building lots of different portals to try and share that data with lots of different researchers or clinicians. So Zero Dash we'repretty proud of. This is a platform we spent a couple of years uh building and we won the good design award about a month ago uh for this piece of software which helps us interpret and translate all of this information. Um so the challenge that we had is our scientists had to look at you know more than 20 different systems to try and work out whether a genetic change is important what it actually means for the patient. It takes us about 96 hours of manual labor from start to end. So this is ripe for being um automated as much as possible potentially the potential benefits of AI in this space. We have some machine learning tools whichhelp us get through this. And so through zero dash we've been able to really rapidly reduce and have the amount of time that it takes. And the less time we can take here, the more time we can spend thinking about patients where the data doesn't make sense. Um yeah, so it'sobviously beautiful and it gives the geneticist the information that they need at their fingertips. And importantly, um one of the biggest bottlenecks is writing the molecular report to summarize. It sounds simple, but if an oncologist is the only person who can write that report, there's another massive bottleneck that you need to be able to overcome. Um so we've been able to try and tackle that part of the process as well with zero dash. Um, all right. We got a couple of minutes, so I'm going to touch I'm going to zoom past this one, but this is the challenge of how do you share genomic data with other research institutes andtry and essentially build the Expedia model where you try and find patients with certain characteristics from a single portal. Um, we've been trying to do this with a national basis with lots of collaborators, but I' I've just run out of time to tell you a little bit about that story. I just wanted to end with I'vegiven you some anecdotes. So, I've got some unpublished data that I probably I couldn't show you today about on the thousand patients how much this is starting to change those survival curves. This is really working and hopefully that'll get accepted in a medical journal pretty soon. This is a brand new drug. Each of these patients has been the blue ones are still on this brand new drug. A bit like that I told you about Jack with the melanoma drug is another type of new drug to target those genes. Um, so I just want to leave you with the optimistic message that medical research is driving these innovations. The pharmaceutical industry is developing new compounds with the research. We run clinical trials to work out which patients will respond to those new drugs. So that this great cycle of medical research and new therapies isthe way that we're going to tackle this disease. Um so I've told you a lot about cancer diagnosis um finding cancer treatment guidance options but genomics has a role to play ina huge amount of human health. Now, from couples, pre-expectant couples doing carrier screening all the way over there. Um, many moms now have a blood test called NIPT to see if their child has Down syndrome. Um, we can adapt that test for cancer detection all the way through to molecular autopsy for a young person who suddenly passes away in their sleep, often caused by a heart condition that they didn't know or a genetic heart condition. So, genomics has got a huge role to play throughout um throughout our life. So, wanted to end on this. Um, this is our fundraiser. It's called CEO dare a to cure. We all get dared to do scary things. So, this is me with a shark. This is a Voytech with a snake uh in Melbourne last week. And so, I just wanted to shout out therelationship that we have with NetApp. You're the third highest fundraiser for our this charity. So, I just really want to thank you fordoing that and your support has been wonderful at this trip. So take-home messages, zero childhood cancer, it's in the name. That's what we're trying to achieve for children with cancer in Australia. Getting New Zealand on board very soon as well. And if that starts to become successful, then I maybe we can start to expand beyond that in our region. Hopefully, you've got theoverwhelmingly optimistic message that precision medicine isstarting to work. It's starting to close the gap on these kids that didn't have any other option. In case you're wondering, there are similar programs in adults with cancer, but maybe not. So not with whole genome sequencing everybody uh but there is genomic smaller genomic tests that are being applied and it's also driving bigchanges in adults with cancer too. Uh I've shown you a little bit about the data ecosystem to support this but so there's challenges and opportunities ahead. We we've we have not solved this at all. Um real world data integration of health data imaging data uh which we don't have yet. How we can bring the genomics into that piece. Um, we're constantly improving how we process our cloud data. Uh, the last talk was all about sustainability. I will admit that we don't look at our carbon footprint right now. We're a little bit focused on this goal, but it's increasingly important. Um, can we identify and predict poor responders of treatment early? Can we detect them failing to respond at day 30? And then we' be able to change their treatment much faster rather than just waiting for them to relapse.um I skipped the global datasharing piece and yeah we're starting to use AI to try and overcome some of the bottlenecks in here. So big generational opportunity to end this disease. Um so just two quick thank yous. Thanks again to my team. Thanks to the zero childhood cancer group, the phenomenal group of almost 80 people now around Australia who are helping to achieve this and it couldn't be done withouta huge amount of funding. So and thanks everyone for coming lace in the conference. It's a really I appreciate it.