Episode #2: The present and future of genomic screening with Dr. Michael Murray

Foresight Medicine Episode #2 TranscriptRobert Rogers: I’m Robert Rogers, host of the Foresight Medicine podcast at the Foresight Medicine Substack, where we are envisioning the future of preventive healthcare as a systematic and whole-body framework for leveraging new technologies to maintain health for as long as possible. In this podcast series, I interview leading experts at the forefront of prevention and early intervention across medical specialties.I am very honored to have as my guest today Dr. Michael Murray. Dr. Murray is Professor of Medicine at the Icahn School of Medicine at Mount Sinai, where he is Chief of the Division of Genomic Medicine and Clinical Director of the Institute for Genomic Health. He is a foremost expert in and a true pioneer of the field of genomic screening and integrating the burgeoning science of genomics into actual clinical practice. In addition to being a leading researcher in applied genomics, he has shaped the field as an architect of genomic education for medical professionals, journal editor, a key contributor to task force guidelines and policy statements in the area of genomic medicine, and as founding director of Mount Sinai’s Genomic Health Clinic. I also want to add that Dr. Murray is the senior author of a beautiful review of this topic, entitled DNA-Based Population Screening for Adults, published January 27th in the New England Journal of Medicine Evidence, which frames the topic we are discussing today, and I encourage all of our listeners to read it, and hopefully our conversation can lead to even greater understanding of this topic.Listeners should note that we are recording this on Wednesday, February 11th, 2026, and this conversation is for general information purposes only and does not constitute individual medical advice. Michael, welcome, really excited to have you on and to be talking with you today.So, one of the many reasons I’m very, excited to chat with you and have you as a very early guest in this podcast series is that the goal of this whole project is really to envision the future of preventive health across different organ systems and categories of disease. And I believe that advances in genomic screening are the foundation of this whole enterprise. It’s really the general-purpose technology that’s opening up new opportunities to identify disease risk at a personalized level. So, let’s start with a pretty general question. What is genomic screening, or as you’ve recently termed it, DNA-based population screening for adults, and how did you get drawn to the field?Michael Murray: Yeah, so, so we’re trying to get people to adopt, the term DNAPS, D-N-A-P-S, just because saying the longer things is a mouthful. But, I got interested in screening, in, the early 2000s. I had left, internal medicine infectious disease practice to become a, a geneticist and refocus my career, and I was at Brigham and Women’s Hospital, and taking care of adults, with rare genetic disorders, but particularly interested in, in bringing the information we were learning to the, to the greater population. And at that point, it was, it was theoretical. There were a few companies that launched in the early 2000s, Navigenics and, 23andMe, and they were starting to, to think about this from a business perspective, but it was really the ClinSeq project at the NIH in 2007-2008, led by Les Biesecker, that was the first to, to do this at scale. They came out with a paper showing that they were identifying cases of BRCA in patients that had nono personal or family history, that would go along with that, and, got very interested in their paper, and left Harvard a couple years later to go to a cornfield in Pennsylvania, where Geisinger Health System is, to really, launch a large-scale project in screening.Robert Rogers: Very interesting on multiple levels, including how you, early on, saw the power of this field and sort of reoriented your career. Just for our listeners, BRCA is a gene that confers increased risk for multiple cancers, and something that actually has come up on another one of our podcasts when we talk about multi-cancer early detection. So. Nowadays, that’s… it’s actually quite striking to think that these key efforts that enabled this new field are only about 15 to 20 years old. But nowadays, when somebody thinks about DNA-based population screening, tell us what… what are the tests? What is the tests that are available? What is the actual information, the genes and the variants that they contain?Michael Murray: So, there’s a couple different technical processes that can be used to create the dataset for an individual. The two, most common are what’s called a whole exome sequence, or a whole genome sequence. And the reason why, projects are deciding one versus the other is still mostly around cost. A whole genome sequence gives you all the information in between the genes. You have 20,000 genes, and they only make up about 1-2% of all the data. Everything else is in between the genes, and a lot of that is regulatory information, regulating the genes to turn on, turn off. Turn up, turn down. And so with, with an exome, or a whole exome, you can just get the 1% to 2% that has the gene data, which is where most of current efforts are focused. You can do the whole, genome and get all the data, but right now, you basically ignore 98 or so percent of that data, just because we don’t know what to do with it.Robert Rogers: And, at your own clinic, can you tell us a little bit about what’s actually offered?Michael Murray: Sure, so, so at the Genomic Health Clinic, we, we have people that come in, sometimes referred by providers for an indication, meaning that, you know, they, they have a family member with a genetic problem, and they want to be tested for it. So we call that, diagnostic care or indicated care. Others come in for screening. They, some are, very educated about it. You know, we have scientists and geneticists that come in, and doctors, and others just have this correct notion that they could learn more about their health risks if somebody looked at their DNA and screened it for common risks. So, we… the most common test we do there is called, well, it’s a 163 gene panel. I mentioned that we have 20,000, and it just picks out 163 genes that are the ones we know the most about, and that we can do something with the information if we find a problem. So, about 80% of those 163 genes are either for cancer risk or heart disease risk. That’s really where we have the biggest area of knowledge. And then the third category is just miscellaneous. There’s a number of different, conditions that we know enough about to look for a problem and then address it if we, if we find a problem.Robert Rogers: Yeah, so I really want to double-click on that, because I think that’s so interesting. As you said, there are about 20,000 human genes, and the report that a patient gets back, or that someone who’s interested in screening gets back, contains information on 163 of them. So, could you say a little bit more about the criteria that are used to determine if a gene and a variant should be included in the report to patients?Michael Murray: Yeah, so the, you know, the, the data, that can be generated through these processes, doesn’t change. This is called our germline data, so it’s different than some of the other conversations you’re having about, about cancer genes that have, have a mutation. So this is the germline data in an individual, and that can be looked at to see if there is a disease-associated or a pathogenic change that can be recognized within the code of the gene of interest. So, I’ll borrow the, explanation from my genetic counselor, who gives a great, description of what we do. So, a gene is like a paragraph of, letters, and what we do is go through and do a spell check, and we might find a change that changes the meaning or breaks the function of the gene. Or we might find a benign change, meaning, the example I always give is, like, spelling the word gray, G-R-A-Y, versus G-R-E-Y. It’s a different spelling, but it sounds the same and it means the same thing. We’ve got lots of those changes in our genes, and it’s the ones that actually break the function of the gene that we’re interested in. Because then we can predict the risk that’s associated with not having that gene function.Robert Rogers: Right, so there’s at least two important elements, it sounds like, to deciding the genes that comprise the test. One is being able to assess the pathogenicity of a given variant, and another is how actionable, or not actionable that is. And, let’s… talk a little bit more about both of those aspects. So, first of all, in terms of determining the significance or the pathogenicity of a variant. So, the DNA sequencing machine, it reads all the letters of someone’s DNA, but then to determine if that is relevant to health or associated with the disease, we need to have a system for describing that variant, and you’ve written a lot about this. Can you describe a little bit about the system that’s used to grade variants in terms of being benign or pathogenic?Michael Murray: Yeah, so in, about 15 years ago, there were a number of diagnostic labs, and they had their own internal databases that they used, and sort of one of the reasons to use one lab versus another is whoever had the most data about what were the problematic changes in a particular gene. So you had… you had labs built just to look at one gene, and they were the experts in that. As the data set started growing and the interest started growing, it was recognized that there needed to be a common repository for that information. So, rules were set up on what causes a change to be pathogenic, or likely pathogenic, so disease associated. What information you need to categorize something as benign or likely benign, so just, a incidental change that doesn’t matter. And then there’s a vast middle where, we don’t have enough data to interpret a gene change. Those are called variants of unknown significance. And if we sequenced, 100 people, we’d probably find variants of unknown significance in common genes in at least, 10 or 20% of them. So it’s… it’s… we’re still in the days when, when we have data that we just can’t interpret. And so, we don’t focus on that when we’re doing screening, because giving somebody back information that you can’t interpret isn’t very useful. For a screening effort, we focus on the pathogenic and the likely pathogenic, because those are the ones that we know are disease-associated, and there’s something to do about it.Robert Rogers: And as you said, you know, if you take 100 people, somewhere between 10 and 20% of them might have a variant of uncertain significance in a common gene. Where does that compare to, say, 5 years ago, and where do you think that will be 5 years from now?Michael Murray: Yeah, great question. So, the, the repository, at one point, I, I did a little, It…day-by-day analysis, this is probably 5 or 10 years ago, in the, looking at the BRCA genes, to see how quickly the information was moving. And there was a pathogenic, variant per day being, being, put into the repository. So, that gene, because it’s probably the most tested gene clinically. That data set is growing, and continues to grow on a daily basis. Other genes, for conditions that are, that get less attention or are less common, they grow much more slowly. And the way that the data grows generally is through clinical experience. So if we have a family with a genetic condition, and we figure out, the change in the gene that’s causing that condition in that family, then we… we get enough information and work together with the lab to, to classify their variant as pathogenic. So it’s… it’s sort of families and individuals where variants are seen multiple times that starts to give us confidence that a gene change is pathogenic. There are some genes where there can be a laboratory test where you can figure out the genes broken when you look at it in, in a cell culture or in a, in an experimental way. So data comes that way, too. But it’s a slow-moving task. There’s a lot of people that are interested in, supercharging this by coming up with better systems for moving something from unknown to either benign or pathogenic.Robert Rogers: So, let’s switch to the other half of the equation of what determines whether a gene is sort of included in the report that a patient might see, and that’s its actionability. And I’m quite curious about this topic, because to me, defining actionability is not so straightforward, and I’m curious, sort of, what the current criteria is, and your thinking, but also going forward, how do you think we can account for variation in patients’ preferences, and perhaps an individualized desire to tailor the amount of information to their own notion of what actionability means?Michael Murray: Yeah, so, I don’t like the term actionability, but I don’t have a better one. Okay. So, I think of it more, being more close, to what, what people have called clinical utility, so…Robert Rogers: I will update to using clinical utility.Michael Murray: Yeah, no, actionability is stuck, and we live with it, but it’s a little vague, and so clinical utility in my mind, and I’m sure there’s a bunch of different, definitions out there, but that’s the idea that, you know, if I screen you and find a change in a gene of interest that I can then offer you, some action or set of actions, whether they’re changes in behavior, or medicine, or surgery, something I can offer you that will either prevent you from getting the problem or attenuate the problem if it occurs. So I’m sort of, on the more classical idea about what actionability is, or what you do with results. For a long time now, there’s been a lot of people that have said, well who are you, or who is the medical establishment to tell me whether it’s actionable or not? What if I want to know the information in order to plan my retirement years, or just make personal decisions? And people often call that personal utility instead of clinical utility. And those are obviously important things, but those are also measurable things, so we could set up studies and designs to, you know, kind of gather the information. If we give somebody a result back on whether they might get a disease or not, and they’re less depressed, or they, or they make life changes that are important. You know, that’s… I would consider that clinical utility, too. But, you know, when we’re in these early days and we’re deciding what to prioritize, I, I don’t see that rising to the top of the list of things that we, we really need to do first, when you figure… You know, we could be preventing, heart disease or cancer in a large, portion of the population. That seems, that from a medical point of view, is probably, the top-tier set of tasks, and, would… one of the interesting things about genomic data is once you have the… once you have the data set, you can use it for all kinds of things. So we’ve been thus far in a… in a world in which the data sets are expensive, especially to do them at scale for a group, and so we’ve focused on these, classical clinical utility idea. But, we right now, in our clinic, run a genome or an exome on people, and, they often ask for their data back. And I always encourage that, at least to save. You never know. But one of the questions I always ask, just out of curiosity, is what do you think.Robert Rogers: And just so I understand, when you say they ask for their data back, meaning they ask for data beyond the 163 genes that are included in the report?Michael Murray: Yeah, so I failed to mention that we do the 163 gene most commonly. But we also do whole exome and whole genome for people, and we get a report that we give them back that has, you know, genetic risks. It matches that 163 gene panel, and then it goes beyond it to talk about things like carrier status for recessive conditions, which have no direct, effect on the person’s health, but… but is… they’re important in, family health and reproductive health,Robert Rogers: We’ll go to carrier stuff in a little bit. We can bypass that for now, yeah.Michael Murray: All right. Yeah, so, when we do the whole exome or the whole genome, people have much more data than has been interpreted, and some of them want a copy of it. So, we work with a laboratory that will, give them their entire file. And so you know, I think over time, more and more people will want that, and more and more people will have options to do something with that data, all kinds of things. Ancestry is a simple thing to get to. We know people have been doing that with genetics for a while now. But also, you know, even making new discoveries that science and medicine haven’t made, I think will have, sort of, people looking at their own data and making observations that then become important for everybody to know about, not just them and their family.Robert Rogers: Yeah, and that kind of leads into another thing that I think is quite interesting, and something that I know you’ve given a lot of thought to, which is how our notion of what is clinically useful, or what has clinical utility, will continue to increase over time, for at least two reasons. One is that as we obtain more genomic information, the number of variants that we learn this might actually be important, this might be pathogenic, will increase, and will decrease the space of those that are of unknown significance, but also because there’s new treatments and new therapies being developed all of the time, and what a few years ago completely lacked clinical utility might today have a clinical trial, and tomorrow have an approved therapy. So what’s our current framework for reassessing information and expanding the notion of what’s clinically useful in a genetic report?Michael Murray: So, a couple ways to think about that. One is, in the, in the large, cohort, projects that I’ve been involved in, there, they’ve all been associated with a research element, so there are people that get access to the data, and participants agree to this. Their identities aren’t revealed, the data is not shared beyond the team. : But they’re going into the data and looking for, new, new things in the data that are important, related to human health. So there’s that research going on. And the idea is that some of that research that might come out of a big project might be important enough that, that we bring it out of the research realm and give it to the patient, to then take to their, doctors and get treated. So the idea that you could, participate in one of these projects, and it might, help us all learn, and then bring important information back to you is possible. The other less cutting edge, but just as important, way in which reviewing the data periodically can help is, for the reasons that you said, that we’re learning about new variants in common genes, that if I analyze your data now in February of 2026, a year from now, two years from now what we called, unknown significance might now be known to be significant, and we might give that back to you to let you go get targeted care.Robert Rogers: It’s also interesting to think about the different care system frameworks that have been set up to do DNA-based population screening, you’ve written about how there’s at least three kind of major paradigms. You can think of the public health paradigm, where a governmental or health authority of a country or a region will offer DNA-based population screening the way that there are blood pressure screening fairs, right? It’s a public health good. There’s the one that I think people are actually most familiar with, which is direct-to-consumer. 23andMe, which started with a heavy ancestry focus, but now many companies say, you know, swab your cheek and send this off, and then we’ll send you the report. And then there’s really integrating it into a healthcare system and care delivery the way that you do at the Genomic health clinic. Did those three approaches tend to have significant differences in the type of information that they are including.Michael Murray: Yeah, so the, the public health idea. A lot of people that, are familiar with that on the healthcare delivery side are most familiar with the newborn screening approach. So, every baby, when born, gets a small blood sample taken and looks for up to 40 different conditions that they might have, many of them, genetic conditions, and so that evolved slowly from the 1960s till now, and it found its way into state departments of public health in the United States. So each state kind of, operates that program, within the United States. In other countries, it’s done differently. Some, it’s through the central government. To do that now, is something that people have discussed, but I’m not sure that, we’ll align all the right parties to… to make that a publicly funded project, in the near term in the U.S. On the other hand, Australia has, has at least done this at at an initial scale, a public health funding and organizing of genomic screening, and they just recently published their data, so a lot of people are watching that closely to see how that goes. The direct-to-consumer idea is now, you know, at least 25 years old with the genetics. And the worry there, has not been, so much about the deliverable back to the patient, but more whether that information that goes back to the patient ever makes its way into the health system to deliver the good that, was… is promised by… by the results.So, 23andMe, in a blog post said that they, they have 20,000 people with BRCA results. And… It’s not clear how often that gets back to, you know, people take it to their doctor, and sometimes their doctor’s an expert in it, they know exactly what to do. Other times, they’re just not familiar enough with it to kind of take whatever the next steps are. So the concern there is that that extra step might be a barrier to optimal treatment. And then the, you know, the integrated programs that we have, I think, you know, this idea that an entire health system would be bought in, and they’d create a process for doing this, is probably the model that, we’ll end up being the most used, unless there’s a change in, in the funding structure, because right now, those health systems are motivated to do extra care for their patient, and right now this is considered extra care. And so, we have a growing number of health systems that are supporting this, but it’s still an expensive thing for a health system to take on, and there’s not really a reimbursement model that works yet.Robert Rogers: Yeah, that is a major challenge, and I would love to touch on that a bit at the end. But since you brought up healthcare systems concerns, I’m curious what your assessment is of how prepared the healthcare system is to monitor people who have been identified as increased genetic risk for a disease. You talked about the issue of sort of inaction around BRCA findings, but I imagine among genomic findings, the hereditary cancer syndromes have a somewhat established pathway for care in terms of increased screening and patients for whom even prophylactic surgeries are offered. But outside of cancer, if you go to your cardiologist, or your nephrologist or your hepatologist and said, I did a genomic screen and I have increased risk for this, how prepared are they to monitor you on a long-term basis?Michael Murray: So as you might guess, it’s all over the map. There’s some people that are very prepared, some people not prepared at all. The, the thing that, that cancer screening has, evolved to, provide is there is a, a national organization, NCCN, that, creates guidelines for cancer management. It also creates guidelines for cancer screening. And so they get very detailed and frequently updated, sort of step-by-step processes for what to do if somebody has a positive screening test for a cancer risk. So in a lot of ways, they are leading the way in how we’re going to need to do it as this becomes more and more common. There are some areas that it might be very valuable to get information back, but it’s still going to be hard to find the specialists that have, sort of, the pathway that needs to be followed. So we’re right… right now working on, the American College of Medical Genetics recommendations for population screening. They should come out later this year, but one of the questions that we’re, we’re really, digging in on is, if we support a list of things to be screened, they not only have to have clinical utility or actionability, but the health system has to be ready to handle this at scale. So if you have a genetic risk that requires a follow-up of a colonoscopy, systems are set up to increase their numbers of colonoscopies per year. But if you have a health risk that requires a subspecialty cardiologist, an electrophysiologist to handle your risk for an irregular heartbeat, then the health system may not have the workforce standing ready to expand in that way, so… so we have to make sure that anything that we’re screening for, we have systems in place and workflows in place that can manage the follow-up, or else there’s no point in doing the screening, to find something that people wouldn’t be able to get the right next steps for.Robert Rogers: Got it, got it. Yeah, that seems like it’s going to be one of the major challenges, but also opportunities for the healthcare system over the next decade or so, and that’s one of the things we’re really interested in exploring as part of this Foresight Medicine project. So, there’s one more key concept that I think’s important for our listeners to understand when we talk about DNA-based population screening, and that’s the concept of penetrance. And for most of these diseases that you would report back in a genetic test, it’s not as if having the variant of concern leads to a 100% deterministic certainty that you are going to get the disease, right? There’s a degree of uncertainty, and maybe just talk through how you think about that concept, and more importantly, explain it to patients, and how they receive that concept.Michael Murray: Yeah, so… so penetrance and variance of unknown significance are things I think about all the time, because they’re the trouble spots. Yeah. So, penetrance, you know, the idea that I have a risk, but might never have the bad outcome associated with that risk is something that we have to continue to work on and address. In my talks, I often show a picture of a 100-year-old woman, with a cigarette that she’s lighting off her 100th birthday cake. And, she… her name is Winnie Langley, and she was… she made the press about 10, 20 years ago, because she was still smoking at 100, and I always… show that slide right after I introduced the idea of penetrance, because within healthcare, we’ve been telling… making recommendations in every area of healthcare, some of which we identify risks that don’t ever become relevant for that person. So I always say that Winnie Langley probably outlived most of the doctors that told her to stop smoking, it’ll shorten her lifespan.Robert Rogers: Almost certainly, yeah.Michael Murray: But on a serious note, the thing that we have to do is if we’re giving people back risk. We have to help them to understand what’s going to happen next, and how we’re going to manage the situation if they get evaluated and they don’t have the disease that we say they’re at risk for. And it’s going to be different in almost every gene, certainly in every category of genes. So, BRCA testing is now 30 years old, and what’s happened there is that periodic evaluations, mammograms and MRIs are done to address the risk of breast cancer in someone that has a BRCA risk variant. It’s not a one-time-and-done situation, and 30% or more of women that find out they have that risk will never get breast cancer. And so we have to prepare people for that situation and help them to understand that we have a plan. If the risk is for something like alpha-1 antitrypsin, we know that there are specific environmental triggers that can be avoided. And so, obviously, coaching them towards that, because that’ll increase the penetrance of the, of the gene risk if they’re smoking. There’s not a whole lot of those very specific, sort of, do’s and don’ts, once we give a monogenic gene risk, because these monogenic single-gene situations are very highly penetrant, that some of the lifestyle measures that we’re used to for lots of risks might not impact it greatly. So we have to help people to understand, you know, you’re going to need to undergo periodic screening, or this is where the data’s at with each situation. One of the things that’s been interesting in doing screening is we can identify a risk in an individual.I’ve seen a bunch of cases of this, and that individual who’s participated in screening in the clinic or screening in a big project does not have the disease or condition that they’re at risk for. But what we always encourage is what’s called cascade screening, so screening out to their brothers and sisters, their children, potentially their parents. And we, we know that if it’s an autosomal dominant condition, that 50% of those siblings or children will have the same genetic risk, and we find lots of situations where the first person doesn’t have the disease, but their sister might, or their brother might, or one of their children might. So, why is it penetrant in the brother and not the original person? We don’t have a deep enough understanding of the biology to answer that right now. But we know that it happens. So you can, from screening, identify a risk that’s actually a family risk that might bring benefit to your family But maybe not directly a health benefit to you.Robert Rogers: Yeah, that’s a really interesting side part of this whole effort, this notion of cascade testing, and who is the patient. It’s not only necessarily the patient who first comes to you, but the larger family. And along those lines, I’m curious if there are big lessons that we can draw from the large-scale efforts that have been done to date, so I think you wrote that there’s something like 5 million or so people around the world who have now participated in large-scale DNA-based population screening efforts, either through a large healthcare system or a regional health authority type of program. And when you take a step back, are there kind of common lessons and themes that you can say, we’ve really learned that we can do these things well from the study of that many people, and other things where you say, these are really the big unanswered questions that still remain after having, processed that many people through those sorts of programs.Michael Murray: Yeah, so, we have a lot of programs going on, but there’s no, sort of common list of ways to do it. So, in a sense, each one of them is learning things that, that could become valuable lessons for other programs. So one of the things that we need is we need to get, these programs to… tell the world about their, their lessons learned as they go through it. Both genetic lessons, as well as just operational lessons about how to do this at scale. So the thing that we don’t have is a common set of rules, and it’s because, all these different programs acting in good faith are doing the best that they can. They’re learning sometimes from things that are published, other times they’re learning from experiences of rolling out screening for other things within a health system. So the gathering of the evidence for how to do this best is something that, we’re probably going to need organizational help on, meaning that, you know an authority like what was created in the U.S. government, to start to advise the 50 different newborn screening programs around the country what a standard practice should be and what conditions they should look for as a minimum list and how to do it. We have to create an infrastructure that’ll do that for population screening with DNA.Robert Rogers: One thing that I think is kind of interesting that’s come out of that work is some reasonable estimates of just the global prevalence of serious monogenic conditions that we can pick up through DNA-based population screening, but it looks like for the cardiac types of things and the cancer syndromes, as many as 1-2% of people are carrying these variants that put them at greatly increased risk, and if we expand our purview to some other diseases and other organ systems, I imagine it would double or more from there. And so, it’s a pretty significant chunk of our societal burden of disease, actually, that’s lingering there. Collectively, these things are… even though each individual one is quite uncommon, collectively, they’re… they’re not rare.Michael Murray: Right, yeah, just the, there’s what’s called the CDC Tier 1, the Center for Disease Control did some work, 15 years ago looking at conditions that have a genetic basis that are common enough that if you screened for them at a population level, you could bring a population benefit. And, those conditions are driven by just 9 genes. If you looked at those 9 genes across every population that’s been looked at so far, so it’s not just certain ethnicities or ancestries, but every population seems to be 1 in 75. That number keeps coming up. Sometimes it’s 1 at 70, sometimes 90, but yeah, it seems that way. And, you know, when that data first came out, Geisinger was one of the first places to have the data, and we let people know that, and people said, well, you know, the population you serve is almost exclusively European ancestry in that part of Pennsylvania. And will this work in other places? And we didn’t have a good answer for that, but now there’s been, large projects that have included, significant, number of people with, Mexican ancestry. In New York, it’s been done with people of all kinds of ancestry, and the number keeps coming up at about 1 in 75. So, that’s a lot of people.Robert Rogers: So I want to switch gears just slightly, although we did allude to this earlier in our conversation. So, we’ve been focusing on identifying and talking about genetic variants that directly affect the health of the person who gets the test. But DNA-based population screening can, of course, also be used to identify genetic carrier status. And for our listeners, I think when many people think of a, quote, genetic disease, they’re actually thinking about these sorts of recessive conditions that are sort of the… what comes to mind when you think of genetic diseases. Cystic fibrosis, sickle cell disease, where people almost never have a, quote, family history of it, because by definition, your parent… one parent carries one mutation, your other parent carries the other mutation, and then statistically, one out of four of their children will be affected by the actual disease. And then there’s also, I suppose, the X-linked recessive diseases. So how is genomic screening handling carriers status?Michael Murray: So, about 5 years ago, somebody showed me a, a publication that showed, which specialist within healthcare are doing the most genetic tests? And, you know before, they showed this in an auditorium full of genetics and public health people, they asked the audience to guess who’s doing the most, and people guessed geneticists, they were wrong. They guessed oncologists, they were wrong. They guessed cardiologists, they were wrong. The group that’s doing it the most are, obstetrics and gynecology, and it’s because now, 7 or 8 years ago, the professional group, in OB-GYN they told their, their professional audience that carrier screening should be routinely offered to all women who are pregnant or considering pregnancy. And so what’s evolved very quickly over a 5-10 year period is this is routinely offered in the obstetrical setting, and the way that it goes forward is there’s a full sequencing of different numbers of genes by different labs, but about 700 genes, that are associated with autosomal recessive risk. And they’re analyzed, and, if the, prospective mother is, positive for recessive risk, then her, partner is, offered the test, and if the prospective father is positive for the same thing as the prospective mother, then there’s a 1 in 4 chance of a recessive disease occurring in any child that they parent together. And so, that’s full-on screening that’s being done in a specific setting for a specific reason.It’s really the example of one of the ways that I think this will grow in other areas of medicine. So that’s being done there. In our screening setting, for instance, in our genomic health setting, we give back the, all the recessive risks to people that get that whole genome or whole exome test, so they get that extra test. And often they get back a short list of, of risk factors that they don’t necessarily have a lot to act on. Many of them have already had their children, but we always encourage them to pass that on to their children, because when they start thinking about having a family, they’ll be of specific benefit. And so, recessive screening right now is mostly in the reproductive area, and it’s moving fast. And it’s really set up a model for just frontline clinicians ordering the test, managing the test, and calling in expert geneticists only when there’s a rare, finding or a problem that, that needs subspecialty attention. I think that’s all medicine will go.Robert Rogers: I’m a little bit curious that, as a system, we’ve settled for the obstetrical office to be the frontline provider there, and not moving it a little bit earlier in people’s lifespan from a public health standpoint. Because if you think about it, genetic carrier screening programs have been around for a long time. I believe that one of the first was for the Tay-Sachs gene in Ashkenazi Jewish communities. That goes back to the mid-1980s. I think, actually, for many years in several Middle Eastern and Mediterranean countries, upon receipt of a marriage license, it’s common to do thalassemia screening, a blood hemoglobin disorder. And to my knowledge, although I’m certainly not an expert here, I think they’ve been relatively successful. Those programs are thought to be successful. And the reason that they were limited to focusing on a single gene that was relevant to the population in which those programs were conducted was mostly one of expense, one of logistics. You couldn’t sequence the whole genome. But now you can, and it just seems to me like it would perhaps be a higher public health priority to move this information to people at an earlier point in their life. Do you think we’re going to move in that direction, or no?Michael Murray: One of the areas of screening that’s moving the fastest is genomic sequencing for newborn screening. So right now, it’s essentially a non-genetic test, the newborn screen. And there’s a lot of conversation I’m in on some of them, it’s not my area of focus, but once you do that genomic screen for a newborn, which is, of course, looking for serious disease in the newborn period, but then that child has that for their entire code, is what are you going to save it for to eventually give back other results that are not related to newborn disease? And one of them is autosomal recessive risk for the next generation. So, I think, we’re limited by, funding and maybe a little bit by imagination about how to do this, best. But as sequencing becomes cheaper and more, sort of, large pilots have been tried for all different, in all different kinds of settings for slightly different reasons. I think and I’ve predicted this in writing, I think what we’ll eventually get to, is that everyone will have their entire genetic code generated at birth, and that will be linked to their electronic medical record. At different points in their life, depending on their gender or their age, it’ll be screened for certain things that are relevant to them at that age and time. And, it’ll also be available to delve into for diagnostic reasons. So,Why don’t we have that yet? Just because nobody’s set up the system and been able to pay for it, but there’s no reason why it couldn’t be. That every child get their entire code, and it follows them throughout their life, and it gets…interrogate it periodically based on either some health concern they have or… or some need to be screened. So, I think we’ll get there. Everybody says, how long? Michael, how long? How long? I don’t know that answer.Robert Rogers: Yeah, well, no, I think that’s a very compelling vision of healthcare delivery going forward. So, just to switch gears once again a little bit, we’ve talked now for over 45 minutes, and we’ve discussed genetic disease really as about monogenic disease, right? But, of course when we think about most of the common human ailments that afflict adults, it’s the combination of the effects of many, many different genes combined with environment, and so that has given rise to a deeper understanding of what we call polygenic risk. Just briefly, talk a little bit about your assessment of the current state of what we call polygenic risk scores, what they are. They have generated a lot of enthusiasm, a little bit of skepticism. I think they’re constantly improving. So, as of now, where do you see their real clinical utility, and where do you see them heading in a few years?Michael Murray: So, polygenic risk scores, like you said, is this idea that instead of finding a single gene that, if not working, puts you at risk for a disease, that you’d look at hundreds or sometimes thousands of small changes across, dozens of genes, that contribute to that risk. One of the things that I think is most exciting about that is that, the way that the results will be reported from those tests. It’s easy for any provider to understand. So right now, one of the factors that’s a barrier to implementation across health systems is that the genetic reports come out, and sometimes you gotta be, gotta have a lot of insider knowledge to even understand what the report is saying. But polygenic risk scores are coming out, and they’re essentially a bell curve that puts that person at the highest percent of risk, the lowest percent of risk, most of us in the big middle for any and every condition. Risk for heart attack, risk for high cholesterol, risk for dementia, risk for less common things like multiple sclerosis or whatever. So, these results will be easily understandable by providers and by patients, and lots of people have started to give them back, and they’re… that barrier of needing insider knowledge is gone. The thing that we don’t have about polygenic risk scores that we need to get is the, clear plan for next steps. So… and that’s going to come through clinical research.So, a lot of people tend to forget that, I keep going back to the example, but it’s the prominent one. BRCA gets discovered in the mid-90s: And, it… there was 10 to 15 years of really intensive research, looking at prophylactic surgery, other interventions to show that they had real value to do those things. So when we come up with the polygenic risk score for breast cancer, and there are examples of that, it doesn’t mean that we can take all the same steps as we do with monogenic risk, because that… that came as a result of, important research, that proved it was valuable in that setting. So one of the stumbling blocks right now from just going out and giving everybody their polygenic risk scores for any condition they’re interested in, is that we don’t have the clear workflow of management that comes from that. So, that’ll come a lot quicker than it has in the past, but we need to work that out for each one. One of the most exciting things in polygenic risk scoring came out last year in the New England Journal of Medicine, it was the Barcode study. This study, took older men, and compared a polygenic risk score to the PSA test, which is a test that looks for increased, risk for prostate cancer, or even early evidence of prostate cancer being present. And everyone in medicine knows that, it’s… It’s the only test we really have for that, but it’s not a great test. There’s a lot of false positives and false negatives. The polygenic risk score outperformed PSA and MRI in screening men for this common cancer, and also in helping their providers know which ones need surgery or chemotherapy versus which ones can be slow-growing and just observed over time. So that kind of, specific use of polygenic risk scores is going to be sort of the leading edge of really getting this into practice, getting specialists and and generalists used to using it, and giving clear benefits. So I’m looking for more examples like that to really be the leading edge for polygenic risk scores to get them into common use.Robert Rogers: Right, so right now it sounds like their actual clinical utility for many conditions is not quite ready for prime time until we’ve defined the follow-up care pathway.Michael Murray: Yeah, and one other thing that I’d throw in is that we also have to prove that it’s giving us something different than what we already know, right? Right. So there’s been,lots of ways to figure out if somebody’s at high risk for a heart attack. If we’re going to add in a polygenic risk score, then there has to be data that it improves the prediction somehow. So, otherwise, why do the test? I think that data’s coming in a lot of important areas, Probably in the next 3 to 5 years.Robert Rogers: So, for the sake of completeness, I just want to make sure that you have a chance to briefly describe, kind of, the one other, or one other type of useful information that comes from genomic tests, and that relates to pharmacology. I don’t want to spend a lot of time on this, but this is how someone’s own individual genetic makeup determines their metabolism of drugs, and that could be very useful information in terms of choosing between classes of drugs, dosing drugs, if somebody needs a specific medication for something that they develop. What’s the current state of pharmacogenomics? How useful is it currently?Michael Murray: Yeah, so great question. Pharmacogenomics is something that’s generated a lot of excitement for at least two decades now, and because it’s so easy to understand, you know, the general population understands that, you know, I took this medicine and I got a better effect than my neighbor, my blood pressure went down, theirs didn’t go down, they had to get a new medication, or…I got a side effect, and they didn’t. So people understand that there’s something about their biology driven by their genetics that’ll cause them to react differently to drugs. So the concept is there, it’s easy to understand. The value of looking for certain gene-drug pairs, as we call them. So, a certain change in a certain gene leads to a different reaction to a drug. The most valuable ones, are limited at this point in time, so there’s a lot of observations on things that while, they, they may have a small effect, don’t have clear clinical utility, to go back to that term. There are a few that that absolutely do. I think the one that people got excited about was the use of this pharmacogenomic approach to decide how to dose warfarin. And a lot of attention and research went into that. But then you know, the pharmaceutical field moved on. They came up with better drugs than warfarin.Robert Rogers: Right, that would have been very useful in 2005, less so today.Michael Murray: Right, and in 1995, even more useful. So there’s a couple really high-value ones right now, and there’s a lot that are, probably useful, and some people, But, but not really driving excitement across healthcare. I think the ones that I look to as being most exciting is there are changes in, in a gene where the gene loses its function, and a drug known as clopidogrel, or Plavix doesn’t become activated in the system, so it’s essentially the same as taking a sugar pill if you have changes in that gene. And then you don’t get the anti-platelet, anti-clotting effect that you’re being given the drug for. So that’s one that’s a clear value in patients that are going to be getting those medications. Another one is there is a gene that can be changed in a way that can cause a very rare, but sometimes fatal reaction to a chemotherapy agent that’s been around for decades called 5FU.So, the difficulty with that situation is that it’s rarely used, and when it’s used, there’s often not enough time to wait around to get a test result for whether you’re going to have a reaction to the drug. Sometimes the chemotherapy gets instituted rather quickly. So we… what we haven’t done is come up with rapid tests for some of these things that we’re going… Until we screen everybody, which is my dream, of course, but until we get to that, we need to have, for pharmacogenomics, we have to have rapid tests that could be done in a time frame before the provider institutes the use of a drug, and that’s been a hobbling, or a barrier to, to, implementation for some of the really interesting and important, gene-drug pairs. The other thing is there’s hundreds, perhaps thousands of drugs that we don’t have that knowledge of pharmacogenomics on. So, we offer pharmacogenomics within our, within our clinic, and, people get back a 32-page report, that, that has all kinds of, important and interesting drug gene effects, but it doesn’t cover every, drug in the, in the pharmacy. So there’s limitations in the knowledge, and then there’s limitations in the implementation, but we need to get there.Robert Rogers: A few rapid-fire questions to bring us home. As we sit here in February of 2026, are there people for whom you would actively advise for or against doing genomic screening?Michael Murray: Screening is always voluntary. We’ll point out to people. You know, if you… if you screen an adult with that 163 gene panel, 4-5% of them will have a positive that they didn’t know about, and they could do something about. So, anybody that’s interested, and unfortunately, anybody that has the you need to have the money to pay for it, because right now, screening is not covered by any insurance, that kind of screening. So, some of the big programs are no cost, so participating in those programs if you’re interested, or getting screened at a clinic like we have if you’re, if you have the money to pay for it. And the cost, a lot of people want to know the number, so that test that I’ve referred to a bunch of times is about $300 out of pocket if you had to pay for, the 163 gene panel right now through a certified clinical lab and delivered through a healthcare system clinic like ours.I’d recommend to anybody that’s, thinking about building a family that they get the recessive carrier screen. About 1 in 50 couples, I’ve read, will get back a screen, where both, both of the prospective parents, have a positive that matches and need to do something about it.Robert Rogers: Is there an age limit past which the utility of genomic screening goes way down? I imagine you’re not gonna… you’re not gonna do much for an 100-year-old. What about 80? What about 60? Where’s the inflection point?Michael Murray: Great question. Honest answer, we don’t… we don’t know. But… Right now, there is data being built around people getting screened when they’re 20 to 30 in the adult realm. Because then you can do a lot of the prevention of cancer and heart disease to really bring measurable benefits across a large group. But there, you know, there are other conditions that are later onset, and so, you could learn things in your 50s or 60s that would still be valuable to you right now. And I imagine, you know, if this, recording lasts 100 years, they’ll be laughing at me for whatever I say, because when everybody’s living to be 130, there’ll be reasons to screen people later in life, too. So, I think, right now, if you said we’re gonna screen everybody in New York City, or everybody in the country. Pick an age that you want to do that at. The number would be 20 to 30 year old.Robert Rogers: That makes a lot of sense.Robert Rogers: People who come to your clinic who are not referred because of some specific medical concern in themselves or their family are not a random sample of the population. These are people who are especially motivated, and I’m just curious, are there common profiles, common motivations that you see in, quote, early adopters of genomic screening.Michael Murray: I think they are people that are, are, motivated to stay healthy and live long. And so, you know, there’s a lot of, different personality types that are motivated by that, but that seems to be the motivation by… by most everyone. There are others, and I think this is important, older adults that are motivated by learning about their risks so that it can benefit their children. A lot of people say, you know, if I find something out here, one of the most important things to me is then to pass that on to my adult children. So I think that’s an important motivation.Robert Rogers: Describe a little bit the patient experience at the Genomic Health Clinic at Mount Sinai. So, genomic information, it really sits at the crossroads of genetics, primary care, multiple specialties. So, how are results typically delivered, and tell us a little bit about the infrastructure that you think is required to do this effectively and responsibly.Michael Murray: So, so we have a, a two-visit, plan for each individual, and that can be telemedicine or face-to-face. The first visit, it surprises some people, but we actually take a detailed family health history as part of that first visit. And the reason is, people come in, sometimes saying, you know, there’s nothing that runs in my family that motivates me to come here, I just, you know, I just want to be screened to be healthy. And yet, when we take a detailed family health history, we can find, reasons to do specific tests in about, 5% of people. 5-10% of people, probably, that come in for screening and don’t.Robert Rogers: That’s a lot.Michael Murray: Yeah, and it’s not that they’re ignoring it, or they’re… or they’re not…healthy, motivated, it just has never been pointed out to them that this story on your dad’s side or your mom’s side, you know, might be important to your health. So, we do that as part of the intake. We talk to them about their current health and their past health, and their goals, and then we spend a fair amount of time talking about the test, what it accomplishes, what it doesn’t accomplish, the limitations of it, because a lot of people think you know, I’m getting my 100,000 mile checkup, I won’t need one for another 50 years. But given the limitations of our knowledge, we tell a lot of people that maybe, in 2 years, when polygenic risk scores are becoming important, or in 5 years or 10 years, they might want to revisit this question again of what can I learn from my genetics? So we always kind of give people that idea. We send it off to a laboratory, we don’t do any of the sequencing at Mount Sinai. Most academic centers don’t do that anymore. There’s reference labs around the country that everybody trusts. And it takes… it’s the slowest test in medicine still. It takes about 2 to 4 weeks to get a result back. And then we schedule a follow-up with the person, no matter what the results are, to, you know, even if there are no findings, we still meet with them to tell them.A lot of times they have the question, well, how is it that I have so many people in my family with X, Y, or Z, and I don’t have the gene that seems to be associated with that? So then we talk about polygenic risk scores and other risks for disease. So we have a good conversation, at the end when we have results in hand. So it’s that two-part interaction.Robert Rogers: Dr. Michael Murray, Chief of the Division of Genomic Medicine at Mount Sinai. It’s been a real privilege to talk to you. You are a pioneer in this field and really pushing it forward, and I’m excited to revisit this conversation in a few years as the field advances. Thanks so much for joining us.Michael Murray: Thanks for having me. Get full access to Foresight Medicine at foresightmedicine.substack.com/subscribe