Scientist Stories: Leroy Hood, Genomics, Systems Biology & 21st Century Medicine
In this lecture Leroy Hood outlines a vision for 21st‑century medicine that moves beyond the reductionist, disease‑centric model of the 20th century toward a systems‑driven, wellness‑focused paradigm. He contrasts the Flexner‑era approach with a new P4 framework—predictive, preventive, personalized, participatory—grounded in massive, longitudinal, multi‑omic data clouds he calls personal dense dynamic data clouds. Hood describes five cutting‑edge, systems‑driven technologies that expand the patient data space: single‑cell analysis of development and cancer, organ‑specific transcript detection in blood, high‑throughput mass‑spectrometry proteomics panels (e.g., lung nodule, preterm birth), peptide‑catalyzed protein capture agents that could replace monoclonal antibodies, and combinatorial synthesis of billions of natural‑product drug candidates in yeast. He argues that these tools enable early‑disease biomarker discovery, exemplified by outlier proteins that flagged impending leukemia and pancreatic cancer. Hood critiques conventional clinical trials as “data‑poor” and proposes a P4 redesign that stratifies participants by dense data clouds to identify responders and non‑responders, allowing two‑stage, smaller trials. He then presents the 108‑pioneer “scientific wellness” study (later expanded to 4,000 participants) that generated dense data clouds—including whole‑genome sequencing, quarterly blood, stool, and wearable metrics—producing thousands of actionable health recommendations with 70 % compliance. Analyses revealed >3,500 cross‑modal correlations, 70 coherent analyte communities, and dozens of drug‑target opportunities. Genetic risk scores for >60 diseases were integrated with these data to predict transitions from wellness to disease, guiding preventive interventions (e.g., cysteine supplementation for high IBD risk). Hood illustrates a preventive medicine pipeline using these data clouds for Alzheimer’s disease, combining early‑stage biomarker detection, multimodal adaptive therapy, and cognitive‑training platforms. Finally, he sketches scaling the approach to hundreds of thousands of individuals for cancer, chronic disease, and broader population health, emphasizing the potential to shift the majority of healthcare spending from disease treatment to scientific wellness.
Chapters
What I'd like to talk about is exactly what Charles was discussing, namely 21st century medicine. And I fundamentally see 21st century medicine as systems-driven thinking about how to deal with wellness and how to deal with disease. And a point I'm going to make in this slide is, ironically enough, Wellness is really going to be the fundamental key to understanding disease. And I got started thinking about systems biology early in my career at Caltech, 1970 or so. And I was really impressed with biology and disease's complexity. And it's very much like the parable of the elephant and the six blind men. At that time, biology was enormously reductionistic, one gene, one protein at a time, or approximately the same. And it's very much like the elephants, the six blind men each feeling a different part and declaring the elephant is a spear or a fan or a trunk, when in fact the elephant is all of those things but very much more. We didn't have a way then for thinking about complexity, and that came to be systems biology. We'll talk a little bit about that. But even more important, to really understand the elephant, you had to make many, many more measurements than the 6 blind men had made. And that's really true for human organisms as well. I mean, if you think about it, a typical visit to a doctor's office, You're lucky if you measure 25 features. And I think what we need to really understand the enormous complexity of humans is billions of features. And that's really what we're going to be talking about. So what's this whole business of 20th and 21st century medicine? So I'll say one, I think virtually all healthcare organizations today are practicing 20th century medicine, unequivocally. And it came about with the Flexner Report in 1910, where Flexner surveyed 155 medical labor trade schools at that time. And he recommended two things. One, you introduce science into treating patients. And two, you introduce science into teaching medical students. The mentoring at that time was virtually nonexistent. So Hopkins took that and really pioneered it for the next 40 years or more. And they used the sciences that were available in the 20th century, the physiology, the pathology, and the like. And the whole focus of 20th century medicine is to find disease and to fix disease. That's the real essence of what it's about. And of course, 21st century medicine is really quite different. And its fundamental feature is it uses systems thinking. And as you'll see, systems thinking has pushed us to three descriptions of 21st century medicine. One is called systems medicine. The second is called P4 healthcare, healthcare that's predictive, preventive, personalized, and participatory. And finally, scientific wellness, and we'll talk briefly about each of those. But these systems approaches use the modern techniques of omics and microbiomes and personalized nutrition, a whole variety of things. But they're all about prediction, prevention, and ultimately pushing the whole compass so that the major resources spent in healthcare will be on wellness. I mean, really what we want is wellness. We don't want disease, and wellness gives you the key for avoiding and/or preventing disease. So the new locus now is what exactly is wellness? And we can define disease with a precision never before imagined, but the really key point is We need to look at transitions from wellness to disease because they're the key to prevention. We'll talk later about that. So what exactly is systems medicine? It's the idea that we apply the principles of systems biology to disease, and that we call systems medicine. And I think there are 3 principal ideas that systems thinking has brought to systems medicine. So one is this idea that in order to understand the complexity of human beings, you need to make enormous numbers of measurements that assay many, many different kinds of systems. And we call these personal dense dynamic data clouds. And these data clouds generate the data that lead for each individual to actionable possibilities to either improve wellness or avoid disease. And what's even more important is this whole approach is precisely what precision medicine should be, and it is not today. It's a far cry from it. But these, as we call them, P4 clouds give us the ability to interrogate wellness and disease in ways we never could before. Of course, the second part of systems biology is the idea that humans have a hierarchical set of biological networks that govern development, physiologic responses, aging, and when disease perturbed, cause disease. So if you can compare the disease perturbed network with its normal counterpart, it gives you insights into mechanisms and it gives you strategies for biomarker discovery and drug target discovery. And you'll see from other speakers actually some examples of this in the future. But one of the other things that systems medicine has really done is it's invented new technologies and strategies that let you explore completely new dimensions of patient data space. And I'm not gonna run through the technologies, and a strategy is one or more technologies embedded in a computational platform so you can get high-throughput measurements. And I'm not gonna run through the strategies, but rather I'll give you some examples of what I think some really intriguing possibilities are. So let me talk about We'll talk about 5 systems-driven technologies and strategies. So one thing that is going to profoundly transform our ability to understand the complexities of human biology and disease is the analysis of single cells. And what this analysis has done, in the case that I've listed here, is allowed us to interrogate the development of iPS cells all the way through to differentiated cardiomyocytes. And we look at single cells at each stage of that development. And what we can map out precisely is a developmental path, and we can see the key transitional points at which you can manipulate the yield of final product of cardiomyocytes. And this takes us into dynamic systems thinking and all sorts of other things where we can integrate enormous numbers of measurements and come up with simple conceptual understandings of major kind of events. And Sui Wong at the institute has used single-cell analysis to interrogate cancer and to discover from a single genome you can create cells with a whole variety of phenotypes, as you can see on the top there, that facilitate various aspects of tumor growth and metastases and things like that. And these insights give us ways to attack these kinds of things that just were not previously available. A second technology that's really interesting is We can look at all the organs in the body and by comparative analyses identify the transcripts that are expressed specifically in one organ, let's say the brain. And then we can ask ourselves, of those, we probably have 400 or 500 brain-specific transcripts, how many actually get expressed in the blood? And what is really intriguing is initially when the blood-brain barrier is intact, very few get expressed, maybe 15 or 20. When you get a disease and the blood-brain barrier starts to break down, you can get hundreds of brain-specific proteins expressed. And what's really interesting is each of those brain-specific proteins is a part have a functional network in the brain, so the ones that are expressed give you insight into the networks that have become disease-perturbed and are manifest in the blood. And we've actually used this to look at Lyme disease in really revolutionary ways and at liver disease too, but it's going to be a very powerful tool for studying the brain. Here's just a representative example of the 200 or so brain-specific proteins that we think can be expressed depending on the state of the blood-brain barrier. A third kind of technology is using mass spectrometry to study the bloods of normal individuals and diseased individuals. And let me say To do this right, because there is enormous signal-to-noise problem, you have to use a lot of systems-driven strategies. And I'm not going to discuss any of those, but rather I'm going to give you the results of what we've been able to do. So number 1, we've been able to get a panel of 2 proteins in the blood that that diagnose with a 50% accuracy whether a given nodule in the lung is neoplastic or is benign. And as you may know, there are probably— there are of the order of 3 million nodules per year that get expressed in the lung. 600,000 of them go to surgery. More than half of them have surgery on benign nodules, increase cost, increase morbidity. And with this simple assay, we save the American healthcare system about $5 billion a year. And it's been commercialized by a company we set up called Integrated Diagnostics. We've done exactly the same for preterm birth, and it's going to have even a more profound impact because The children of preterm births are an enormous cost for society that goes on the entire length of their lifetime. So we have a marker that at 19 weeks can, with enormous accuracy, predict— 19 weeks of pregnancy in the blood can predict who is going to go on to have a preterm baby, and that gives you plenty of time to give them progesterone or other strategies that delay pregnancy to bring the fetus to term and so forth. So it's going to be an enormous savings there. Another technology that is really important has been pioneered by Jim Heath, who has just become president of ISB this past January. And these are peptide-catalyzed protein capture agents. And what he's done is really transformational. He's created a library of, say, 10 to the 6th circular 5-mer d-amino acid peptides, so they're not susceptible to enzymes, enzymatic degradation. And what he does is take the protein you'd like to make the capture agent, around, expose him to this library, and he finds monomers that bind with low affinities. And there's a click chemistry that can join together two such monomers in a way that preserves their three-dimensional orientation on that protein. So what that does is it makes a dimer protein capture agent. And on average, These dimers have picomole sensitivity, which is a sensitivity that's as good as the very best monoclonal antibodies. The— these agents now can be made from peptide epitopes of a protein, and in doing so, you essentially eliminate cross-reactivities, so you can do something with these you'd that you never could do with ordinary antibodies and so forth. So these are obviously terrific reagents to think about diagnosis, either in vitro or in this case in vivo diagnosis. And what we did was had a protein capture agent designed against human PSA that was labeled with a copper atom and did PET scanning to show that incredibly rapidly after injection, the localization of that in vitro diagnostic reagent went right to the prostate. And of course, this opens up the possibility for being able to look at the dynamics of how molecules move around in the course of their action and so forth. So these reagents are truly incredible. They're enormously stable. You could put them in an envelope, send them to Africa, they'd have the same affinity they did before. They are sensitive. We've made trimers and tetramers, and each additional mer we add on gives you close to a log increase in sensitivity. They're digital. Once you know the monomers and the chemistry of the click reagent that links them together, you can put all of those in a test tube with copper catalysis and and make as much of the reagent as you want. So you store how to do it in a computer. It doesn't require living animals or living cells. Another thing that's really key is the minimal cross-reactivity. And what is really cool is you can design these reagents so they can beautifully penetrate cells, either for diagnosis or therapy. We're now starting to create miniaturized ELISA assays which will be able to take a droplet of blood and make all the analyses you need. So for example, we're taking both the lung cancer tumor nodule panel, two different proteins, and we're taking the two proteins in the preterm birth panel and and we're gonna create ELISA assays and it'll be a cheap way to screen millions of women in a very easy manner. In fact, it can be set up to do it in the doctor's office quite easily. We're now just scaling this up so it's large-scale production and so forth. I haven't talked about it, but these are going to be incredibly, effective drugs because you can design them to do things that you want with unbelievable specificity. And all the antibody protein drugs that are present today have real limitations because of their cross-reactivities. So, uh, in vitro, in vivo diagnoses and therapeutic reagents. And my prediction is in a 10 to 15-year period these will completely replace monoclonal antibodies. So we'll see. And the final example I'm going to give you is a beautiful example of how to synthesize and screen rapidly a billion new potential drug entities. And these are natural synthetic products. And I'll say in the— armamentarium of drugs out there today, 70% of the existing drugs are either natural synthetic products or they're derived from them. So it is an incredible space in which to search for drug targets. So the basic idea is he— Mike has devised a method where you can make a billion natural synthetic products very rapidly. And he has the ability to screen them and look either at cells, targets on cells, or targets on proteins for less than 1/10,000th the cost of what a pharma has to do, to give you an idea about the scaling. So the really key point then is DNA synthesis makes it easy to synthesize microbial metabolic enzymes. And what he's done is taken about— he's now got about 2,000 of these metabolic enzymes synthesized, and he orders them together in a combinatorial random manner in cassettes of 4 or 5 or 6. And those are put into back vectors and put into yeast, and each yeast then will synthesize the natural synthetic product specified by the order of those metabolic enzymes. So if you have a million yeast with a million different combinations, you're gonna get a million different drug products, and you can use these products to screen use high-throughput screening techniques for yeast, and that's why this is so inexpensive, and he's already synthesized lots and lots of products. And the really cool thing is this is going to be the solution to drug resistance in hospitals because you'll be able to use this to turn around the new drug that deals with the resistance. Very rapidly and very effectively. So those are all techniques that are systems-driven, that are exploring new dimensions of patient data space, new kinds of opportunities. So P4 medicine and healthcare really is the convergence of these 5 things. So systems medicine, scientific wellness, digital health, all the devices we know about that make various physiologic measurements, big data and its analytics, and finally social network. And what the convergence of these things has done is let us specify with considerably more detail the systems-driven features of P4 medicine, ultimately P4 healthcare. So how do the two compare, P4 and contemporary healthcare? Well, P4 is proactive and contemporary medicine is reactive. P4 focuses on the individual. Contemporary medicine derives its major actionable possibilities from populations of patients that average, unfortunately, the properties. And we'll talk more about that in just a second. It focuses on wellness as well as disease, whereas contemporary medicine is all about disease. And it is all about using these dense data clouds I spoke of earlier to assess environmental and genetic contributions to each individual. And of course, contemporary medicine does not use that approach. But one of the things it does that's really radical is it says, clinical trials are done completely wrong today. So what do you do for a typical cancer drug? You'll take 20,000 patients, you'll give a set of them placebo, a set of them the drug, and you'll take the data from that large set. But what you realize is the fundamental assumption in that clinical trial is that all the individuals are essentially identical. And that is totally wrong. They're different genetically. They're different in their lifestyle and environmental exposures. What P4 medicine does is say we'll have 20,000 individuals each with their data cloud and then we'll stratify those individuals according to key questions. For example, who responds to the respond to the drug and who doesn't respond to the drug. And with the data clouds, you instantaneously have assays that can distinguish responders from non-responders. And why— and there are other kinds of questions we could ask, too. And why is this so important? These are the 10 top-selling drugs in the U.S. today. The orange individual is the responder and the blue individual is the The blue individual are the non-responder. The best result is 1 in 4. The worst result is 1 in 25. Look at the money that's been wasted and the morbidity that has arisen from that. When you do a traditional clinical trial, you have no idea who responders and non-responders are. The traditional clinical trials are data poor. And when you do the P4 approach to it, you immediately have the responders and non-responders. So that leads to the possibility of doing a small initial trial, getting the markers for responders and non-responders, and then doing again another small trial with all responders. And we'll talk about that new strategy for drugs and so forth in the future. So for P4 Health, it's about wellness, it's about disease. 96% of society's resources are spent on disease, almost nothing is today spent on wellness. And we've actually designated something we call scientific or quantitative wellness, and we're gonna talk about that in just a moment. And what I would say is in a 10 to 15 year period, My prediction is scientific wellness will be a whole discrete sector of the healthcare agenda, and its market cap will far exceed that of the disease sector of healthcare as society appropriately readjusts its resource allocation from disease toward wellness. Now, it's really important, therefore, to think about wellness And in 2014, we decided that ISB, Nathan Price, and myself, that we would persuade 108 of our friends to go through a 9-month period of generating these dense, dynamic, personalized data clouds to see what that led to in two regards. One, did we really improve the health of individuals? And two, How profoundly transformational were those data? And the answer was very positive in both regards. And these dense, dynamic, personalized data clouds that we're talking about here are really— I liken them to the Hubble telescope because they give us the ability to look at human biology and disease with a resolution that's never ever before been possible. And we had a large paper in Nature Biotechnology in August of 2017, and we drew 3 conclusions, and I'm going to talk a little bit more about these later. We can do incredible correlations. We can actually determine your genetic risk for more than 60 diseases, and we beautifully see wellness to disease transitions for many different diseases. We'll definitely come back to that. What are the data clouds? They encompass the complete genome sequence. We took blood samples every 3 months so we could get clinical chemistries and metabolites and proteins. We took fecal samples every 3 months to quantify the gut microbiome with regard to different species, and then we used a Fitbit and other measurements to do quanti-self, and that led to data for each individual that when analyzed led for each individual to a long list of actionable possibilities that if acted upon could be used to optimize wellness or avoid disease. Everybody had so many actionable possibilities that what we did was hired coaches that every month delivered a few actionable possibilities to the individuals, one, explaining what they were in simple terms, and two, all of them were trained in psychology to persuade people to make appropriate lifetime changes. And that led to a 70% compliance with the actionable possibilities. Which was really spectacular. One of the most astounding results from a personal point of view was all of these well people realized they weren't nearly as well as they thought they were. And in fact, if you had this staircase of wellness, I'd argue that the average person is pretty close to the bottom. And of course, what elevates you up those stairs are these actionable possibilities that we discussed. And a really key point is as the data increases, the number of actionable possibilities increase exponentially. So there are essentially an infinite number of ways you can improve your wellness. And what we really aspire to is to take individuals all the way up this ladder to where Eric Topol's Welterly are. So, these are people in their 80s and 90s that go into their 90s mentally alert, physically fit. And we think almost all people can be taken into that range. And what is really interesting, as you pass into your hundreds, people die incredibly suddenly. Then from complete systems failures. So the aspiration is to be able to get people into their 90s healthy, hopefully wealthy, 'cause you won't have spent so much money on healthcare and everything. And then when you get into the hundreds, you're on your own. You have to deal with systems failures. And of course this means that scientific knowledge should be a lifelong journey. So many of the people were really enthusiastic about continuing this program that we started a company called Aerovail in mid-2015 that took 80% of the pioneers and then started delivering to consumers, and again, we call them pioneers, the benefits of scientific actionability. And we're now in a place where we really see that within a 3-year period we can have 100,000 people with 100,000 data clouds that we can analyze and keep. And we think by 5 years we could well be approaching half a million. And it might go up actually much faster than that depending on the cost of the assays and things like that. So what Arivale has been able to do already with the 4,000 individuals they have are see wellness to disease transitions for virtually all the major types of chronic diseases. And keep that in mind because we'll tell you why that transition is so important. As I said before, we published 108 Pioneer data in August. And actually, that's the statistical correlation on the COVID of the issue. And the statistical correlations are really spectacular. So we could look among the 108 and identify 3,500 correlations where we asked, Does a data bit in one of the 6 types of data correlate with any data bits in any of the other 5 types of data? And again, 3,500 positive or negative correlations. And these were multiple hypotheses tested. And now we have 4,000 people and the statistics are really strong and all the things I'll talk about here are generalizable to that point. But that's an incredible number of correlations. And when you look at them as an intact blob, it looks exactly like a hairball. So what we did— it's not quite what we did, but it's close to explaining it— is to take that hairball and begin cutting edges that represented the lowest statistical associations. And what that cutting allowed us to do is reveal communities of analytes that were clustered together, and all of those communities correlated with biology, correlated with genetic disease risk, correlated with biomarkers and a whole variety of things. We found 70 of those communities. And what the communities can do is really pretty incredible. That's one community that we're bringing out, and there you can see all of the associated outlines. And that has to do with LDL cholesterol. And the interesting observation there is that thyroxine has a negative statistical association with LDL cholesterol. And of course, it is now a drug that's used to fight and lower LDL cholesterol. We found in those 70 communities 9 other examples of drugs that have already been identified. We probably have 200 examples of drugs that remain yet to be discovered. So an enormous opportunity for drug companies in the future. The second thing is using genome-wide association data together with your complete genome sequence, we can determine your genetic risk for more than 60 different diseases by taking the GWAS SNPs, positive or negative in nature, and mapping them as probabilities into your genome, and that gives you a number that we can then take 2,000 normal genomes and get a bell-shaped distribution, and then we can show where you fall on that distribution to say, are you higher, lower, medium risk, and so forth. And this risk is important in two ways. One, if you're, for example, at a very high risk for Alzheimer's, we can follow you really closely And we know some of the early markers to say you're beginning the transition, so you can be treated immediately. But we can do that for the other 60 diseases as well. And number 2, I don't know how many of you are going to understand this, but this is really an exciting idea. So bear with me. So the interesting thing that we can do now for all of the diseases that you see there, those are the GWAS-associated diseases, is we took the 108 pioneers and we put them in 5 bins from lowest risk to highest risk. And you see that on the slide behind me. Then what we did was we compared the risk level against all of the analytes we'd examined in this analysis. And we found in the case of every single genetic disease, there were striking positive correlations and negative correlations with genetic risk. Now, all of these people are healthy. None of them have cardiovascular disease. None of them are on statins or any of those kind of things. If you take the ones on statins, you get an enormous bent at the top. But what that means is that this LDL that we see correlating in a beautiful linear fashion, we know how we deal with it now. If it's very high, one of the things you can do is be given statins. That really brings it down. But operationally what that does is it changes your genetic risk for this disease phenotype. So all I'm saying is there are 59— Hey, could you go outside the room please if you're gonna talk? Whoever it is. All I'm saying is there are 59 other diseases, and let me show you another one. So this is a correlation of increasing genetic risk from left to right for inflammatory bowel disease, and the analyte cysteine goes down in a dramatic way. So what about these high-risk people? If we gave them cysteine, would that change the probability of getting disease? And cysteine's increase makes a lot of sense because it's on the pathway that goes to glutathione, and glutathione is key to reducing oxidative damage that is one of the hallmarks of IBD. So absolutely a logical thing. So this is something we can test, but there are many, many other things we can test in this same way. Okay, disease state transitions. Here's an example of a woman or a man, I can't remember which one this was, that the Aerofail coach noted their lymphocyte level was enormously, enormously increased. And in fact, at several different blood draws, he was advised to go in and see his doctor. And finally, on the third one, he did, and indeed he had leukemia. But what was interesting is we had 3 bloods prior to the point of diagnosis. And what we did then was to look at all of the other people in the ARAFEL group, and it's of the order of 1,000 at that point in time, and map every single analyte with regard to looking for outliers, okay? So what we found here is with 1,000 people, There was one protein that was way, way out, an outlier. There were two other individuals that had that protein elevated but weren't outliers. But the interesting question is, are they on their way to becoming leukemic? So we can follow them and study that. That protein has to do with stimulating B and T cell function. Is known to increase the risk for lymphoid and non-lymphoid malignancies and so forth. It's an incredible potential biomarker and/or a drug target. So we have to do tests to see which it is. And in a similar vein, we had a woman, a 57-year-old woman in January of 2017 get diagnosed with stage 4 pancreatic cancer. We had 4 prior bloods. We looked at those bloods again with regard to each of the 1,000 or more analytes, and we found one incredibly intriguing protein that is out there on the outlier stage. And to a first approximation, as you get closer to diagnosis, the level of the protein goes up. So this protein is a notch receptor. It's primarily found in the beta cells of the pancreas. It's an incredible target either as a biomarker and/or a drug. And again, when we have 100,000 individuals, we'll be able to compare many, many people to see whether in fact these indeed are biomarkers. So let me talk about preventive medicine of the 21st century, and it has has to do with the transitions we've talked about. As I already said, with 4,000 people, we're beginning to see many, many wellness to disease transitions. So the simple idea is that we use the data clouds to get biomarkers to mark the transition, and then we use that to interrogate the transition with regard to a whole series of these systems approaches. We ran through so that we can identify candidate drugs and/or therapies that could reverse the disease at its earliest stage before it ever manifests itself as an irreversible disease phenotype. We think we can do that for most chronic diseases, and I'll show you one that I think we'll be doing it for in 5 years or less. And that's Alzheimer's disease, we'll end up with that. But this is going to be the preventive medicine of the 21st century. We spend 86% of our healthcare budget on chronic diseases. Suppose we can eliminate most of those in a 10 or 15 or a 20-year period, that's going to transform cost of healthcare. The PD-3 clouds, these data clouds that we've talked about, have many applications. You can optimize wellness. We've shown now beautifully you can use them to study the progression of disease and the response of disease to therapy. And what's really interesting is this two-stage approach to clinical trials. So in the future, what we're getting set up to do is have an initial clinical trial with 50 patients. Drug and placebo, and you look for 3 things. One, stratify the disease into its different major subtypes. Two, identify biomarkers to distinguish responders and non-responders. And three, you look if there are off-target hits, toxicity. Then the drug company with this 50-person study can say, do we want to go on to the second one? And the second one is 50 patients, all responders, they'll get a 96 or 98% result, and the FDA will approve that number, that drug with that number of people. Genentech did exactly that with Herceptin 10 or 12 years ago. So we think this is going to really revolutionize clinical trials. I haven't time to talk about it, but every single experiment we do now is an N-of-1 experiment. Experiment, and it's really the key to deconvoluting complexity in so many interesting ways. And this is really exciting— these NF1 experiments now are beginning to give us the metrics we'll need to quantify wellness and to assess the rate of aging. And we could talk about that, but we don't have time to talk about it here. So here's a systems approach to a complex disease, and it's Alzheimer's. So point number 1, of course, is the history of Alzheimer's is grim. In the last 12 years, there have been 400 clinical trials, zero successes. In the last 15 years, there has not been one new Alzheimer's drug that's been discovered. Last month, Pfizer announced It was getting out of the Alzheimer's business because it's too complicated. And the basic problem is it's a complex disease and no single drug is ever going to really touch it. So you have to go to multimode therapy. So what we're doing is we're going to use a consortium of outstanding people who've contributed the three fundamental paradigm changes we need to really deal with Alzheimer's. So one is this idea of the data clouds that you can use to assess transitions, response to therapy, all these other different kinds of things. The second idea is something that Dale Bredesen at UCLA has pioneered. In fact, there's a book on the New York Times bestseller list called The End of Alzheimer's. And what he's pioneered is a multimodal, individually adaptive therapy that has 36 points in the regimen. And what is striking is he's used this on about 200 early-stage Alzheimer's patients, and 85% of them have responded really well. So he thinks he can reverse from early diagnosis back to normal. In fact, he's done experiments where the person went back to normal, went off the regimen, went back to Alzheimer's, went on the regimen, went back to normal. So it's a pretty convincing demonstration that what he's doing is really working. And the final one is something that's been pioneered on brain health by Mike Mrazanek at UCSF. And basically what he spent 20 years doing is demonstrating that you can computationally assess and train people toward brain health. And he's shown beautifully the assessments get you really early stage Alzheimer's diagnostics. So we're putting all of these things together in various combinations. We're going to take high-risk Alzheimer's patients and follow them with the cloud because we want to get the very earliest stage of transition. What is classically used now is a psychological diagnostic it's called mild cognitive decline. And with PET scanning, they've shown there are changes that occur 6 to 8 years before you ever see, so you're well into the disease by the time you have mild cognitive decline. That was the test that Trump actually passed in his physical examination. So what we wanna do is, there is a website that has 3,000 individuals with really high genetic risk. And we've begun exploring people on that website and we're going to put them in the PD3 cloud study over an extended period of time to be able to gauge the very earliest transition and to get biomarkers from it. And there are lots of other things we're going to do. Large population, long-term studies to look at these transitions with, starting with the classic MDC, mild cognitive decline things. And we'll be doing cognitive assessments to look at the early diagnostic markers as well. The therapy is going to be a mixture of this multimode therapy of Dale and this cognitive training of Mike, which I think is going to be very powerful. We have clinical trials in Alzheimer's in a very preliminary way going on at Providence now that are testing both of those types of therapies, not together yet, but we'll do that in the next stage. We think all of this can be scaled to the entire US. The cognitive therapy is easy because it's computer-based and We can imagine putting it on your cell phone and just using the clicks you use in the ordinary business of your life to assess where you stand with regard to cognitive health and so forth. And we also can imagine that we can get these earliest markers and we can put these peptide-based antibodies to make ELISA assays for from a microfluidic format from a drop of blood and have absolutely standard large-scale screening. And then from those two screens, we'll funnel these people in an appropriate manner into the therapies and so forth. And the vision is, A, to get rid of Alzheimer's. B, we think it's going to work on all other neurodegenerative diseases. We think this is absolutely a generalized strategy for complex diseases. And all of them are going to require multiple multimodal therapies. Very rare is going to be the disease that will let you cure it with a single drug. And ultimately, what we'd really love to do is bring brain health to the world. So some examples of what we can begin to think about doing with systems approaches to cancer. Suppose we have this population of— well, 500,000 should be 5 years. It's a typo. But suppose we have that many people. We have all of these transitions. We have the ability to get the markers. We have the ability to do N-of-1 studies in a combinatorial manner with appropriate kind of drugs. I mean, I think cancer— certainly falls in the category of needing to have multiple drugs to do it. So the diagnosis, the PD-3 clouds you've seen have really given us transition points. The— what we can do now that I think is really exciting is we can look at patients before and after they've transitioned into disease and with 100,000, we'll get a lot of those for any given cancer. And we can do the correlations and we can look at the communities and we can see how the communities have changed when you transition from wellness to disease. That gets you into biology, biomarkers, drugs, and of course the outliers are biomarker drugs and they can even hint at mechanisms. Prediction and prevention. Identify the earliest transitions and reverse, as we said earlier. Identify high-risk populations and follow for the earliest transitions. The therapy is going to be multimodal correlations. Again, in exactly the same kind of way as you begin to give them drugs, you can see how normal, undrugged, and the drugged people have changed in their statistical correlations and their communities of analytes that reflect the fundamental changes in biology that have occurred. The two-step clinical trials, I think, are gonna be key. And one thing we're doing with Providence too is we're using wellness to take cancer patients that have been through the rigors of therapy and bring them rapidly back to wellness. We're doing that for breast cancer patients. Be starting 100 patients hopefully this year on just exactly that kind of trial. So here are the people that did the 100 Pioneer Wellness Project. Many of them went right straight to the company, Arivale, and set it up and got it going very rapidly. And there are lots of people in the combined lab of Nathan Price, myself, and a number of them have made contributions. So thank you. And I'll answer questions, although I suspect there's no time left. So who's the guardian of the time?
More from Joshua
Scientist Stories: Hans Waterham, The value of phenotypic assays in peroxisomal disorders
Scientist Stories: Joel Zlotogora, Genetic carrier screening programs for reproductive purposes