WEBINAR TRANSCRIPT

LAUREN: Hi, everyone, and welcome to Rare Perspectives, a new online roundtable series that brings together innovators, influencers, and thought leaders in the rare disease community to discuss the unique challenges of diagnosing, treating, and developing therapeutics for rare genetic diseases. My name is Lauren Chunn, and I’m the editorial director for Rare Perspectives. I’m excited to have you join this roundtable discussion on the end of VUS and the impact on the diagnosis and treatment of rare diseases.
Before we get started, I want to encourage you to participate in the discussion. You can submit any questions you have for the panelists in the Q&A panel below at any time during the webinar, and we will answer those questions during the Q&A session as time permits. While we are introducing our panelists, two poll questions will pop up on your screen, and the results of the “Will 2030 be the end of the VUS?” poll will be shared after introductions. We will also relaunch this poll near the end of our webinar to see how opinions might have shifted as we discuss. The final results will be shared in a follow-up email after the event. Please note, as well, that this webinar is being recorded, and you will receive a link to the on-demand recording within 48 hours.
Now, I’d like to introduce our panelists. Today, we have with us Dr. Heather McLaughlin, the Senior Director of Molecular Diagnostics at Pharming; Dr. Nicole Miller, who is the Vice President of Molecular Diagnostics at Ultragenyx; and Dr. Mark Kiel, the Chief Scientific Officer at Genomenon. I’d like to give a moment for each of our panelists now to tell you a bit about their background and why they are passionate about rare diseases. We will start with Heather.

HEATHER: Hi, everybody. I’m Heather McLaughlin. I’m the Senior Director of Molecular Diagnostics at Pharming. I am a clinical molecular geneticist by training, and very recently, I’ve switched over to the pharma side. My focus really is on molecular diagnostics, making sure that patients have access to the very important genetic testing that can diagnose their conditions, and elucidating the genetic, clinical, and mechanistic landscapes of rare disease. VUSs are very near and dear to my heart, and I think it’s very important that we continue along this effort to resolve them. I’m just really excited to be a part of this conversation today.

LAUREN: Great, and Nicole?

NICOLE: Yes, good morning, everyone. Thank you so much for the invitation to participate. I’m Nicole Miller. I’m Vice President of Molecular Diagnostics at Ultragenyx, and we focus on ensuring a timely and accurate diagnosis for rare disease patients. That can occur through many different ways. Ensuring that VUSs are resolved in a timely manner is one of the primary ways that we can ensure rare disease patients get a timely and accurate diagnosis.

LAUREN: And Mark?

MARK: Yeah, thank you, Lauren, and thank you, panelists and audience members alike! I’m Mark. I’m the Chief Science Officer and founder of Genomenon. My background is as a molecular genetic pathologist, but I also have a background as a patient, and I have a personal story of how I benefited greatly from efforts by many in the medical community and surrounding to help patients like me get diagnoses. In that spirit, I founded Genomenon to visit that benefit onto other patients as well. At Genomenon, our mission is to save and improve lives by making genomic information actionable. I think this is such a relevant topic to talk about: variants of uncertain significance and how we can as a group make a great effort toward reducing the burden of VUS.
It’s no coincidence that this webinar is being hosted the day before Rare Disease Day, fittingly on a leap day. It underscores the sheer number that rare disease patients have in their collective. There are 300 million rare disease patients worldwide, and each of them deserves greater awareness to strive for improved diagnostics and better treatments. Today, in our discussion with Nicole and Heather, we’ll be focusing on variants of uncertain significance and how they can drive enhanced diagnostics.
I mentioned Genomenon’s mission of making this information actionable in the service of saving and improving lives. We do that in two different ways, both for rare genetic disease as well as cancer, helping increase diagnostic yields. We also work closely in drug development, helping design precision therapeutics based on that evidence, targeting the molecular drivers of these diseases, and further, to leverage the sheer amount of this data to help unlock new DNA sequencing applications to better understand and intervene for these rare diseases.

LAUREN: Great, thank you, Mark. And now I will leave it to you to get the conversation started.

MARK: All right, thank you, Lauren. To kick off the conversation, many of you may be aware that the NHGRI, the National Human Genome Research Institute, came out several years ago now, I think almost a decade ago, with a number of bold predictions for the field. Among those, arguably one of the more important ones, is the end of the variant of uncertain significance, which is the scourge of diagnostics, a challenge for clinicians, and a cross to bear for patients as well. Heidi Rehm and Douglas Fowler followed up recently with an op-ed about some optimism about getting closer to that goal of ending VUS. We’ve discussed that internally at Genomenon. I’m pretty sure that it’s a topic on lots of individuals’ minds within the community, and we thought that we would formalize that discussion here with my august panelists.
VUS delay diagnoses, they delay treatment, they’re consternating to physicians, and they’re confusing to patients. For variant analysts, whose task it is to reconcile these VUS, it can be very challenging to sift through all of that evidence and make informed decisions. With that background in mind, the context that has been set by leaders in our field, let’s begin, panelists, by talking about strategies, tactics, and trends in the service of reducing the burden of VUS.
Heather, I’ll begin with you. In your experience and in your discussions, what are some of those strategies, tactics, and trends that you’re seeing that can help eliminate VUS?

HEATHER: Sure. We have the more traditional means, if we’re talking about a child, we can test the parents. Was this variant inherited by an affected parent or an unaffected parent, or did it occur de novo? We can do segregation analyses within the family. But when we’re talking about resolving VUS en masse, we really need to understand the molecular mechanisms of disease. Why do mutations in this gene ultimately lead to disease? Is it because there’s a gain of function there, or there’s a loss of function there, or there’s a binding domain that’s very important? Really, the key to really being able to do this on a gene-by-gene basis is to very deeply understand exactly how these mutations ultimately lead to the disease within the cell. Then we can use that information that we have — known pathogenic variants, known benign variants — to sort of make inferences on what novel variants of uncertain significance may do.

MARK: That’s great. Nicole?

NICOLE: Resolving VUSs are so important for rare disease patients because, when you go and get a molecular diagnosis for the condition that you’re displaying, I would say, at least in our experience with ultra-rare diseases, we’re seeing that about half of the time, the patients have a VUS. If it’s autosomal recessive, they may have two or one VUSs, and that makes their diagnostic odyssey extend. It’s really important for these VUSs to become public and for the clinical data and evidence, biochemical evidence, to come associated with those VUSs and for families to know that there are other families that have that same VUS through their physician. It’s incredibly important to resolve these to end this diagnostic odyssey.
If you look at patients with rare diseases, they can spend six years in a diagnostic odyssey, seeing over 17 different clinicians when they’re trying to reach their diagnosis. For us, one of the ways that we try to end this odyssey is by publishing all of the variants that we have seen associated with rare disease. We do more than the public databases do. We also show the genotype-phenotype relationship. We show, broadly, the number of individuals that have been associated with that, that are unrelated, if we can. That’s some of the approaches that we take for VUS resolution.

MARK: I’m sensing a dichotomy that converges on the clinical, the data, the idiosyncrasies of the disease, and making sure that the evidence is appropriately adjudicated and published, but then also research. We can treat those differentially. Sort of naively, you could wonder if we’ll just eventually have sequenced enough patients that we will have seen everything, and we will know everything. Nicole, to your point, as long as it’s available publicly, individuals who see a patient with a variant are much, much more likely to see it in the public databases. I’ve recently given talks where the evidence suggests that VUSs are not decreasing; they’re increasing. This may actually be the opposite trend that we’re facing. I’d invite you, Heather, to weigh in first. You can take either the clinical side or the research side, because both are important. Which do you think is going to be more strategically valuable, or do you think they’re going to be equally necessary in this quest?

HEATHER: Yes, absolutely, both are going to be necessary, right? The one thing that happens with rare disorders is that often, when they are initially published in the literature, what we see are families that are more on the severe spectrum of disease. Then, as we start to sequence more and more individuals, we start to understand, my goodness, there’s a huge clinical spectrum of this disease. That really helps us understand the mechanisms of this as well, but then, on the other side, the mechanistic side, we start to understand that variants in this particular domain tend to lead, from a molecular standpoint, they lead to this type of cellular phenotype, and maybe variants in this domain lead to a different cellular phenotype. When we take all of those together, what we’ll find is that the research efforts that are undertaken really allow us to better understand those clinical differences that we start seeing in patients. Hopefully, as we resolve all these VUSs, we’re not only going to resolve them, but we’re going to gain a much deeper understanding of the disease itself.

MARK: Yeah, Heather, you speak to scale. There’s scale on both sides of that fence. There’s scale in terms of how many different groups are working at larger scaling sequencing initiatives, some many involved with screening, which is very foundational to our understanding of genetics and how it has bearing on human disease. But then, there’s also the ability to massively parallel research initiatives for understanding the molecular mechanisms. Nicole, you can choose either; you can do both if you’d like. Which of those two do you think is going to be the most meritorious?

NICOLE: Well, I’m just going to respond to what you were saying in terms of the VUSs increasing. I just want to make a comment on that. I think a lot of times, we see quotes, and they’re designed to be provocative. “VUSs are increasing, not decreasing.” We have to look at the real application of sequencing, genetic testing, in the context of clinical presentation, and then I’ll link it to research. There’s a couple of elements here. One is, are you talking about the gene associated with the disease that the patient? I don’t think that’s what they’re really referring to there when they’re saying that VUSs increase. What’s happening is, if you go beyond that, and you sequence the genome, when you’re looking not just at one gene but the whole genome, sure, you’re going to find a bunch of VUSs because you don’t have the clinical phenotype associated with what would help you classify that variant. That’s one comment.
The second is, a lot of this is a function of numbers. How many genomes have we sequenced? We could probably mathematically predict how many we have to sequence from how many different people from different parts of the world until we can really understand if we’re resolving VUSs. Right now, if you’re of European descent, I would say, yeah, you’re on the side of a better chance of getting a rare disease diagnosis than you are if you are not of European descent, because our reference genomes are not fully representing all of the different backgrounds of people.

MARK: That brings up an interesting point about there being a sort of hierarchy of VUS. I say, not all VUS are created equal. Many are the result of large-scale healthy individual screens. There are some artifacts of sequencing that could populate as VUS, there are obvious polymorphisms, but then, you said they’re sequenced in individuals that have no phenotype. There’s that need of, not just publishing all of the variants in their number, but also the evidence and the origin of those VUS, so that the clinician who sees a patient can better adjudicate the meaningfulness of the result in the public databases or in the published literature.
I know that many of us in the community are associated with newborn sequencing initiatives, which is an adjunct to newborn screening by phenotype. This is sequencing for genotype. What a richness of evidence and information, especially longitudinal, to better understand, better predict the genotype-phenotype correlations, and how to intervene. Do we have any optimism there about all of that collective data being put to the service of reducing VUS? Heather?

HEATHER: Yes, so newborn screening data, or newborn screening in general, I think, is really part of the call to fight for this resolution of VUSs. Newborn screening can’t ever be fully effective until we are able to definitively determine at the time of sequencing what these variants mean. The really interesting thing about genetic newborn screening being more available to folks is that it really allows us the opportunity to understand from a genotype-first approach, will this individual eventually develop disease? The important thing to remember there is that we’re not necessarily expecting that individual to be born with the disorder in question. Many of these disorders are progressive in nature, and we might expect them over their lifetime to have disease. That’s another reason why it’s really important that we really understand those molecular mechanisms of disorders. In many cases, there might be some biomarkers we might be able to use to predict if this individual is starting to show signs, maybe even cellular signs of disease, before the overt clinical manifestations of disease occur. I’m very excited about the potential of newborn screening to give us much more information about not only the variants but also the clinical spectrum of disease.

MARK: Nicole?

NICOLE: It’s an area I’m very passionate about. Newborn sequencing does bring a lot of promise. It gives us an opportunity to better understand the human genome, because we will be expanding beyond just European ancestry genomes. What Heather said is extremely important to call out, that you need additional evidence in some cases when you have, for example, a VUS or a variant without full disease presentation, because at birth, you may not have a strong set of clinical signs and symptoms alone that help you diagnose, and so you may need additional testing. It’s interesting because some of these newborn sequencing efforts are not even including VUSs; they’re studying but not including. That actually, in some regards, is going to hurt patients more, which is why a lot of these efforts are going to come out in parallel to the standard biochemical method, which we know works well.
I think it’s going to be, to Heather’s point, essential in many cases to do the biochemical testing to confirm a diagnosis. I’ll say something just to be provocative here. I would ask us to think about, is a molecular result in the newborn screening period an actual diagnosis? I’ll say, maybe that depends, because we know that even in healthy individuals, you can have, very rarely, a molecular genotype that says you should have disease, but you don’t. I think it behooves us to go further than just the molecular, and we might even want to codify that differently in our medical record, not as a diagnostic result, but as a genetic screening result, just like it is for the newborn period. It’s a newborn screen; it’s not a diagnosis.

MARK: That’s great. I’ve been involved in some of those discussions, not at the center because I’m not a patient-facing clinician, but I’ve heard those caveats, and that’s a very important point to make. I’ve heard people on the pessimistic side of newborn sequencing need to be counseled that this is not a diagnosis; this is a trigger for monitoring, downstream workup. We have to be cautious about the words we use or the inferences we make when we think molecular equals diagnosis. It’s a guide.
I’ve thought about the number of patients that we’re struggling to diagnose because of this VUS challenge. An even bigger challenge I’ve referred to as “occult VUS,” or patients hidden in plain sight, who don’t even have sequencing. We’re not there yet with newborn sequencing at a large population level. Can you help me, both of you, understand ways to increase sequencing of patients who have reason to be sequenced as a way to minimize misdiagnoses and help better understand, better reconcile VUS to disease-causing or benign? Heather, do you have anything to say there?

HEATHER: Sure. First of all, education is a huge component to this. Physicians or healthcare providers have to be comfortable ordering the genetic testing in the first place, and in order to be comfortable, you have to understand the very basic things. What is dominant/recessive, what does heterozygous mean? Then, understanding the framework for how we interpret variants is important as well. I have the privilege of attending conferences for my job, so I interact with a lot of immunologists in my capacity. I’ve been very impressed with that group of physicians and the way that they have embraced genetic testing, because they’re starting to understand that the results of this test can truly change the life of a patient, depending on what we find. It really takes an eagerness to become comfortable with your discomfort in ordering that genetic testing, but also understanding that there are things we have in our toolbox that we can do when we do identify a variant of uncertain significance. We can do clinical correlation, sometimes there are research-based methods we can use to learn more about that variant, we can use family and segregation data.
Trying to break that stigma of genetic testing as this really scary thing, we’re sequencing hundreds of genes, and how am I going to deal with this information? We’re almost in a place where we have a duty for our patients and the quality of their lives to perform this testing and to keep learning more and more. I think, again, being comfortable with being uncomfortable is the first step to really embracing the use of genetic testing and making sure that we’re servicing our patients in a manner that’s going to lead to better health outcomes.

MARK: In my training, two experiences, one on the clinical side, one on the lab side — on the clinical side, don’t order a test if you’re not going to do anything with it. Early days, you didn’t know if there was anything to do, as you said, Heather, especially when clinicians say, “Why do I want to order a 100-gene panel? I don’t know what to do with it.” Well, increasingly, we’re able to do something. We’re better able to understand for diagnostic, prognostic, or therapeutic purposes. Then, on the laboratory side, there’s a need to not proselytize, but to educate the clinicians. You have the test, you have the panels, you have NGS. It was fascinating to me coming up in training to know that that was a challenge, to teach the community, even in a tertiary hospital setting, these tests are valuable, and here’s what they can do for you. Nicole, how would you build on this — or be provocative? I think you’ve been provocative twice already.
NICOLE: [LAUGHTER] That’s possible, again, too. I do think education is a big part of it. VUSs are like the big bad wolf, really, and it makes folks afraid of screening more broadly. My vision is that this upcoming generation is generation G, for genome. Got it, and I really feel like we have to demystify and educate how genetic testing can improve diagnosis of all patients, including rare disease patients. The VUS is getting in our way in some regards. The more we embark on a campaign to educate, the less that will be in our way in terms of using genetic testing. Right now, there’s heterogeneity, for example, in the return of laboratory results. If you have a larger panel, you might not even get a VUS reported, if it’s a single VUS and is in an autosomal recessive gene. That’s unfortunate, because 14% of the time, your second variant isn’t found right now in autosomal recessive disorders, so you might actually be missing a clue that you have a rare disease when you don’t get a full report. On the other side, you can understand why people don’t want that sometimes, because they feel like, as clinicians, how do I talk to my patient when I’m looking at a dozen VUSs from this very large genetic test result? So it’s about learning, resolving VUSs, and education.

MARK: To build on that, I was thinking almost exclusively on the education of the clinical community and in community hospitals, and those who are later in their careers and haven’t gone through the DNA genome revolution. You’re bringing up a great point about patient education. I talked about my experience with foundations, they do a great job of organizing the activities and educating their patients. I’m seeing great trends in patient empowerment, especially in rare disease. It’s not infrequently the case that I know about parents who are better informed about the disease and molecular mechanisms of their kiddos or the patient’s disease. Getting that, it’s a challenge because we’re merging statistics and genetics, and those are very complicated things. When you put those two things together and enfold them in health outcomes, that makes for a difficult educational hurdle to get over.
I want to pick up on something else that you said, Nicole. I’m not a diagnostician, I wonder if either of you are aware, if you have a VUS and a VUS, two VUSs in a gene that causes autosomal recessive disease, is there an art to acting on that as a clinician and proceeding forward even though the variants don’t pass that threshold of being likely pathogenic? There are differential ways to respond to that in cancer and rare disease. Patients and their families, they want answers, they want to be bold. The clinical community has to be cautious and move according to evidence. I wonder if you have any insights or notions to share about the actual diagnosis using VUS.

HEATHER: In the scenario that you described, typically, there are some additional things that you can do to tease those two things apart. My first question anytime I see two VUSs for an autosomal recessive gene is, are these in cis on the same chromosome or in trans on opposite chromosomes? One way we can determine that is testing biological parents. You can also test children, or there’s long-read sequencing that can be utilized sometimes. Answering that question can sometimes give you a little bit more confidence and allow you to rule in or rule out.
Then, I always try to look at what the gene is and what the pathway is. Are there biochemical assays or any other phenotyping assays that can be performed? I work in the immunology space. Can we look at certain cell subtypes? There are certain tests for certain genes that can be ordered that give you more clues and insight into what’s happening with that variant of uncertain significance. I still want to acknowledge the challenge that a physician would face in that scenario. What you’re asking them to do is to take that genetic information, take the clinical information from the patient, and then also take any biochemical or functional data that you have and harmonize and bring all of those things together to make a decision for your patient. That might be the only patient you ever see that has that exact clinical scenario. It’s very difficult for folks to go about that.
One thing I also wanted to mention, that probably Nicole and I should have mentioned in the previous conversation: I don’t want folks to forget about the importance of genetic counselors, too. The importance of genetic counselors, both entering clinics of subspecialists that are just starting to embark on genetic testing, but also, when we’re talking about these family situations, genetic counselors really serve an absolutely critical role in helping families understand their genetic testing results and even helping healthcare providers understand what those genetic testing results mean within the context of a family. Sorry for going back there, but I just didn’t want to miss that opportunity to talk about genetic counselors.

MARK: That’s beautiful. We have a number of genetic counselors on our team. Obviously, they’re working on the data now, but what a noble enterprise, being at the forefront and communicating to patients this complicated information, and helping guide them through the complexities of what’s to come for diagnosis, treatment, and long-term care. Nicole, anything to build on there?

NICOLE: Yeah, no, and thanks, Heather, for bringing up the genetic counselors. I think they’re so essential to rare disease families and all families with genetic disease in understanding their disease. A lot of this is about conversation between the clinician and the laboratory. When you get a VUS result, it is definitely important to reach out to the laboratory and ask about any additional evidence that they may be able to provide, because they may have additional biochemical evidence that wasn’t submitted. Often, when you’re submitting your genetic test requisition, there are some boxes for you to check and some places for you to fill out information that may not be completely filled out. There’s often additional information that can be provided. In association with the genetic counselor, you can also get some more information on segregation. Sometimes you are dealing with someone who’s adopted, though, so you may not be able to get all of that information, but the conversation is still really important.
While I think VUS resolution is going to really skyrocket with all of the new tools that we have, including what you offer at with Mastermind at Genomenon, and through some of the tools you’re going to talk about, there’s still going to be, in my mind, an importance about that clinician, patient, laboratorian, family discussion. It still needs to happen.

MARK: Good. In that context, let’s change topics now to whether VUS will end or significantly taper in six years, by 2030, or if we can extend the goalposts a bit. Is this an unattainable ambition? We talked about some of the tactics. We also alluded to some of the complexities. I want to dig into those complexities and where those tactics are likely to culminate.
I’ll start here, to be a little provocative. I’m going to say the AI words. I’m going to say artificial intelligence. Undoubtedly, folks in attendance here have heard about AlphaMissense. We’re all likely aware of previous computational models that use evolutionary conservation and structural biology and 3D models of molecular folding. There’s a lot. I’ll euphemize it, to say it’s very heterogeneous and it’s of only limited value in its current instantiation. AI has had breathtaking advances in language processing and question answering and, more recently, in video production. We have this great embarrassment of data that we’re sitting on: functional data, clinical data. It’s basically this storehouse of fodder for these AI models. Heather and Nicole, can you speak to what you’ve seen of AlphaMissense and the like, what you’ve discussed, what you think about the potential, where it’s at now, and where it may be in the future?

HEATHER: Sure. When I was in training, I remember a lot of these models up and coming. A lot of times, we would evaluate each and every model, and then we would do projects to go back and look at known pathogenic and known benign variants and determine how accurate these predictions actually are. Unfortunately, at that time, it really wasn’t that great. You really didn’t have a ton of confidence if it said that this is predicted to be damaging, that the variant was actually going to turn out to be pathogenic. I think what has happened is a lot of those models were sort of built on variants at a genome-wide scale, so it was we weren’t looking at a particular gene or even a particular type of variant. It was just, let’s smash things together and then see what happens and we’ll try to predict it later.
What we’re seeing now is, we’re having gene-specific models that are created. We know the different protein domains in this gene. We know about known pathogenic and benign variants in a particular gene. The models are able to get a little bit more sophisticated because of the knowledge that we have about the genes. Many models that we’ll see coming forward are going to be more gene-specific or maybe even pathway-specific models that are going to be much more predictive of whether a variant is pathogenic or benign. Those are what is going to allow us to push things forward.
The other thing that I want to state is that, I don’t think we’ve really talked about ClinVar a ton, but part of the way that these models are effective is that we’re able to take known benign and known pathogenic variants and input them into these models. Hopefully, when you run the model, you should start to see pathogenic variants go one way and benign variants go another, and then there’s a clear demarcation. In order for us to have those variants to train the datasets to create these models, we really do need to continue our data sharing. That’s not just laboratories uploading what their classification of a variant is, but also sharing the information that led to that classification. How many de novo occurrences are there? Is there segregation data? This is all information that is, unfortunately, siloed currently within many different clinical laboratories across the country. I know when I say that, that’s actually a big ask. Some laboratories have evaluated over two million variants, and you cannot expect laboratories to individually submit that type of information for each and every variant. So we really need to find a way to make it easy, and to have a benefit for laboratories to provide this information. Until we do, I don’t think we’re going to have the strength of data that we need to properly train the models.

MARK: That’s a good point. It’s the expert crowdsourced model of data sharing. You’re asking altruists to be even more altruistic, right? They’re already dedicating their lives to patient care, and they’ve got time constraints in the clinic, and they’re doing many things, often in research. This is an added ask, and that has slowed some of the progress that otherwise would be made in making this data public.

HEATHER: Yes, we certainly need to find ways to incentivize that to happen, because just asking for it is sort of unreasonable. I think we need to try to find ways to incentivize laboratories to actually provide that information, and that’s really when we’re going to see the ball moving forward.

MARK: Nicole?

NICOLE: You know, I think that’s such a great point. I also was going to answer along the lines of, I don’t want to say entirely, “garbage in, garbage out,” but the volume of data that comes into the AI models needs to be sufficient in order for us to have confidence. It’s just a numbers game again. We have to have enough data from diverse genomes in order to get a better understanding in association, as Heather was calling out, phenotypes. There’s still this desire to own something unique with respect to genetic information. Laboratories are not necessarily incentivized to put all of that information into public databases. There could be a paradigm shift approach that could help us here. I’m seeing a question in chat about EMR as well. You know, if we think about what’s in the medical record, it’s getting better. We’re now getting ready to ingest genomic data into the medical record, but we need to really think about this now and make a paradigm shift with the data that go into the medical record. What do I mean by that? I mean that we should barcode genetic data with the history of that variant.
This question we see in chat asks, what do you do if you find a variant in a medical record without additional information? Well, optimally, that variant will be in the record with a barcoded set of information that gives you a little more history. How was that variant found? Was it found in what type of sequencing effort? Was it a single gene, was it a panel, was it a genome? What was the methodology? Was there a trio result around it? I could go on, but that that helps also in federation of data. It simplifies federation of data because right now, with all these genomic efforts, everyone’s worried about everyone seeing everything that’s behind their curtain. I don’t think you have to show all of that. You could get to a place where we can move forward more quickly if we share agreed-upon data points about these variants, with respect to clinical information.
The last comment I’ll make about that is we’ve thought about data federation from the context of understanding our genome until now, but we’re now moving into data federation in terms of clinically diagnosing people, and those are different. How you federate data for those purposes, I would offer, is different.

MARK: Yes, it’s not just about the volume of data, it’s about quality. Who’s going to arbitrate the quality of that data? Who’s going to agree on the structure of that data, so that it’s imminently shareable? Who’s going to preserve the sanctity of the patient’s autonomy and anonymity, with ethical and legal and social discussions? As I said, I’m not central to those discussions, but I’ve seen those, heard those played out, and there’s a need to proceed cautiously but also a need to proceed as quickly as possible. So there’s a bit of tension there.
Getting back to this section’s topic, will VUS end? I’ve got some poll results coming in here. I’ll share them at the end. In prep, I know, Heather and Nicole, where you stand. Let me push a little bit and say my concept of VUS is in an arc of history. Nicole, you talked about time and history and documenting the patient data and the functional data that goes along with it. If you think about the past and our knowledge of genetics, every variant was a VUS, because we didn’t know what was going on at all. I think the first one was in sickle cell, when we first decided what genetic mutation was associated with disease, but biology knew. It was known — they were basically VUSs in waiting, waiting for evidence, waiting for insight, waiting for clinical characterization. If we think about this question, will VUS go away by 2030, I think that’s an argument to make: it’s just a matter of time, and the question is, how long will it take? Is that a bad way to think about it? Patients and genetics are complex, and there’s not a one-to-one correlation, “you have this variant, you’re going to get the disease,” because biology is sloppy. Heather, you can pull apart any of that, those pieces.

HEATHER: As I was mentally preparing for this discussion, one thing I kept thinking about — I almost feel like there’s going to be a few phases of resolving some of these VUSs. One thing we’ve been talking a lot about in the community is, this variant of uncertain significance term is really quite a large bucket of variants. We have variants in the middle, where we know absolutely nothing about them, or almost nothing. It’s going to be quite difficult to resolve that type of variant, but we also have variants that are on the other ends of the spectrum. We have variants of uncertain significance that have segregation data, that have pretty strong functional data, and then we have variants where they’re just very, very common compared to the rarity of the disorder, and we’ve got negative functional data.
The middle step of this is that we start to have individual variants of uncertain significance that are maybe leaning pathogenic or leaning benign. Until we can finally get to the point where all VUSs are eradicated, I think the other thing that will happen is that the genes that have been most popularly sequenced and studied, those are the genes that are probably going to achieve this first. The BRCA genes, some of the cardiomyopathy genes, these are genes that have been very well studied. Those ones are probably going to have this happen to them first, which means we still need to continue to fight for those genes and disorders that are considered to be ultra-rare. Those folks have a much higher uphill climb, but I don’t think that it can’t be done, and I know that there’s a lot of folks that are very interested in making that happen. Go ahead, Nicole.

NICOLE: Yeah, I think we’re going to get largely there. I don’t know about all genes by 2030. That’s a little bit of a stretch, but I definitely agreed with what you’re saying, because the biology is complex. I want to just point out that diagnosis is not a molecular diagnosis, it’s a clinical diagnosis. A molecular test or a genetic test is one component of that clinical diagnosis. Physicians should feel emboldened to make the clinical diagnosis, and that molecular result is a tool. We talked about that in the context of newborn sequencing, for example.
For the size of the gene and how often it’s been sequenced, and in how diverse a population, is a function of how quickly we’re going to get to VUS resolution. CFTR for cystic fibrosis is a great example. Early on, there were maybe dozens of variants, and now we’re in the thousands. It would be interesting to hear from that community, if they feel that this gene has been saturated. This is actually a mathematical problem, and it is solvable, but rare disease and human evolution are complex. I don’t think there is a binary answer for diagnosis in every single case.

MARK: I’m being told by my handlers that we have a number of questions. This topic has clearly struck a nerve — very active questions. We have more than a thousand registrants to the webinar in the audience. Let’s close the formal part of the roundtable discussion before we turn over to Q&A with, let’s suppose that there is a future, all the caveats and advisors that have been laid out, where VUS are no longer such a significant problem. What will that mean for patients, for clinicians who treat those patients, and even for folks who are developing the treatments to improve the lives of those patients?

HEATHER: The resolution of all VUS paints a very hopeful future for me. I think we’ll start to see higher reimbursement rates, hopefully for genetic testing in general, and probably more uptake in genetic testing. We talked about providers, and especially in certain non-genetic subspecialties being very uncomfortable with the fact that they may get a VUS back. That takes the piece of ordering out of the equation. For our patients that are navigating a rare disease, it means a shorter time from diagnosis to treatment. We see many of our patients with rare disease having these large diagnostic odysseys. Sometimes their variant was identified when they were quite young, and they lived for 5-10 years before that variant was established to be disease-causing. I’m hoping that time between testing and actual molecular diagnosis can be shortened. Nicole and I have talked about newborn screening in part of this talk too. I think it’ll be very powerful in newborn screening and hopefully finally allow us to realize the full benefit of newborn screening. If we know what all variants mean, we can more accurately predict what that might mean for that patient as they are born.

MARK: Beautiful. Any prognostications, Nicole, for a future without VUS?

NICOLE: A future without VUS would definitely shorten the diagnostic odyssey. This is huge for all people. If we can envision that as our future, this is huge for all people. I think it’s amazing. I also want to just point out, Heather made a great point about VUS resolution, going backwards in time. You may have a result in your medical record that’s a VUS, and not all laboratories go back and will inform you of that change. It’s not just about patients going forward; it’s about patients who’ve already been sequenced. To address this issue, if we do resolve all the VUS, we also have to change our behaviors and make sure that we go back in our record and let those patients and families know that that variant’s classification has changed, which is not uniform today. Not all laboratories do that regularly.

MARK: It’s a challenge, as I mentioned before, about submitting your data publicly. It’s an added burden to an already oversubscribed clinical apparatus, and it’s important. We have to, as you said, change our behaviors to figure out a way to make it happen.
I think, Lauren, you’re going to field the Q&A. If you want to come back, we’ve got some poll results here. Ending on the topic of “will we come close to ending,” I’ll hedge a bit, “VUS by 2030,” the poll results are here from the audience, “yes for all possible VUSs” by not surprisingly, 6%. It is a very ambitious goal. However, “yes, but only for well-characterized,” as Heather talked about, there’s going to be phases or there’s a distribution of what we’re going to be able to attack here. That’s actually 53%, and I hope that they stuck to the letter of the question, “by 2030,” because that’s a very happy outcome. That’s in striking distance, within the four or five years hence. Then, “no, not possible by 2030” is 30%, “not possible at all” is 12%. I’m happy with that skew, a skew toward optimism. It’s a good way to conclude the formal part of the roundtable. Lauren, I’ll turn it over to you for questions from the audience.

LAUREN: Yeah, we’ve had a lot of really great questions. I want to thank everybody for participating and submitting your questions and thoughts. I’m going to try to summarize some of the questions that have repeating themes to see how much we can get through. There were a couple of questions about polygenic effects and how the field is going to address that when it’s not as simple as interpreting one variant alone, and instead thinking about the effects of multiple variants together. Either Heather or Nicole, do you have any thoughts about how the field might be addressing polygenic variants?

HEATHER: Yeah, I can start with that one. I’ve been following pretty closely the reporting of polygenic variants, and especially clinical testing for polygenic variants. I think the elephant in the room there is that, unfortunately, given the data that were utilized to determine the relevance of these polygenic variants, there are certain populations that are not well served by those polygenic marker predictions. European populations, sure, that’s typically what data sets were used to do that. So, I’m excited about it, but I still feel like we have a long way to go, because there’s a disparity in which patients are actually able to access it and utilize that data right now.

NICOLE: I’m going to switch my backdrop now to a rare disease day backdrop, so we can all remember that that’s coming up tomorrow. To answer the polygenic risk score question, I think it’s compelling and yet, I think we have to be incredibly cautious in terms of using polygenic risk scores for the reasons that Heather just raised. Using VUSs just complicates that matter further. Some of what I’ve heard about use of polygenic risk or concerns me when it’s used in absent clinical data, clinical evidence. I think that’s the biggest caution I would mention there. I’ll leave it at that.

LAUREN: Okay, great. Thank you both. One other question about widening the net for VUS searches beyond just coding variants. The question is, “is there still a sense in the field that these are less important, generally?”

HEATHER: This is kind of a similar question. Right now, the great majority of what is sequenced are the coding regions, and we’re not really sequencing promoters and deep intronic regions and other regions of the genome. I don’t think we can really predict whether… We just have not done enough sequencing and testing and functional assessment of those types of variants to really truly understand. I dislike it when people say like, “oh, those don’t matter,” or “variants on another transcript don’t matter.” I don’t really think we know that yet. Let’s use some caution when we’re interpreting them and trying to understand them, but to throw them all out is going to be a mistake.
Some of the folks that voted that VUS would never be all resolved are probably thinking about those types of variants, actually. Those are probably going to be, unfortunately, on the three prime end — I’ll use a little genetic joke — of figuring VUSs out. But I’m hopeful that we’ll get there eventually.

NICOLE: Those are definitely in that 14% of why we aren’t finding all the explanations for genetic disease. I do think that they’re important. What’s coming soon with long-read sequencing should help us quite a lot there. Just improving how we sequence DNA and looking at RNA is going to help. The non-coding variants, I think, are very important.

LAUREN: Great. Some other questions we’ve gotten have been around the concept of sharing VUS results with patients. One is a question of whether VUS should be included in healthy patient screening, for example, newborn screening. Other comments from patients, for example, have shared that that information can be helpful with finding support groups and, preparing to receive new information that might be published at a later time. Either of you have any comments on sharing VUS results and whether we should do that and how we should do it, and so on?

HEATHER: I am definitely in favor of sharing VUS results. I don’t necessarily think it always has to be on that primary clinical report, though. The way a lot of laboratories have approached this, particularly as the use of very large genetic panels and exome and genome sequencing has occurred, is a supplemental table, for example, that can be shared with patients or their provider to have that further information. As we’re learning more and more about genes and as new therapies are becoming available, I do think it’s important for patients to understand, as long as they truly understand that it’s still, at this time, a variant of uncertain significance.
Knowledge is power, and the more information that we have about that variant, what may not be meaningful today, in a very short amount of time, could be quite meaningful to that individual. So, at least having access, or encouraging folks to provide access, especially when requested, is very important. I always think that patients should have access to their own genetic information, especially if they’ve requested it. That can mean access to the genetic testing report all the way to access to their actual sequencing data itself.

NICOLE: Yeah, I think this is a really important question. With newborn sequencing now, it’s a consented activity. It’s not being done as newborn screening occurs. That gives an opportunity for programs to potentially return some of these results, so we can build and learn. Right now, especially when we’re learning with newborn sequencing, it’s very important to understand the history of that patient and what happens over time, and to follow those VUSs. There should be a mechanism, as Heather was saying, for families to know about them. We need to think very carefully about consent, because during this process, you want to encourage people to participate and feel good about participating, but also balance creating too much fear about something that may not come to pass. I think it is important, but it should be done cautiously.

LAUREN: Great. A similar question, more on the physician side, is “how might you suggest we educate the non-genetics workforce to address molecular results in the midst of so many other hats and demands on that workforce?” And similarly, “what tools are available to physicians who find test results, for example, within an EMR that are not explained or explained sufficiently?”

HEATHER: I think all of us can agree that training of the workforce is a very large issue. That goes for molecular laboratory geneticists, for genetic counselors, for physicians, medical geneticists as well. The training of the workforce is going to be extremely important. I’ve been at a few conferences lately where there were entire sessions about VUSs and/or about how to explain genetic test results to a patient, and what to do if you are in a situation where you have a VUS, the patient’s phenotype fits, and you’re very, very suspicious, and what should you do? I think it is going to depend by specialty, the different steps that you can take. It’s very helpful for clinicians who have successfully gone through that process to say, “here was my patient, this is how they presented, I ordered the genetic test, I got this result back, and here are the steps that I took, and the outcome that I got from that.” That can be very powerful for other physicians or clinicians, to see someone else doing it and to feel like they have a good support group.
Also, just eliminating the stigma of genetic testing being scary. A lot of folks are fearful to admit that they may not know everything about this particular topic. The more that we can come together and support each other, it really is only going to benefit the patient.

NICOLE: Yeah, if you have, in your medical record, particularly a VUS, the first thing that I would do would be to contact the laboratory if you have the record associated with that. If not, it actually is worth looking at locus-specific databases, ClinVar, some of the public databases, for sure, and of course, looking at Mastermind with Genomenon. At Ultragenyx, we find that it’s very important, we feel it’s very important to share what we learn with the patient and medical community. We publish all of the variants that we know about, with the appropriate consent to do so, in locus-specific databases. We’ve generated a website called rarediseasegenes.com. For the diseases that we have curated, the gene variants that we’ve curated, you can go and look to see if that variant has been found before. We’re slowly building through those genes on rarediseasegenes.com. Looking for disease-specific, locus-specific databases, for sure you should look at Mastermind, you should look at ClinVar, there’s also LOVD. These are resources. Your big resource is going to be a genetic counselor, going back to Heather’s comment earlier. Genetic counselors are going to also know about these resources and be able to help walk you through it.

LAUREN: Right. Speaking of ClinVar, which you just mentioned, one more question related to that, and that’ll be our last question as we’re coming up on the hour here. “Any comments on the recent massive release of 275 million new variants from All of Us? Was ClinVar ready, and how will so many variants be handled in terms of characterizing them?”

HEATHER: Yeah, first of all, we need to get used to this, right? As Nicole said earlier, this is what we wanted to happen. These sequencing studies have been, as many of them are starting to near completion, they’re getting to a phase where they’re able to actually share this data. Personally, I am very welcoming of any additional datasets, especially those that can inform allele frequencies in populations that have been historically underrepresented in many of these databases. That’s the first thing. We weren’t ready for the genome to be sequenced, and now we’re not ready for this influx of variants. What we need to do is view it as an opportunity to learn more about these genes, learn more about these disorders, and learn more about these variants, and really use it as one powerful tool to hopefully start to resolve some of these VUSs. The more data, the better, but it can feel like you’re drinking out of a water hose, at this point in time, with all the data that are becoming available.

NICOLE: It’s hard. Right now, there’s a gap, often of six months or longer, for variants to make it in ClinVar, even without a high amount of variants being submitted at one time. That’s why when we built our database, we were using our sponsored genetic testing program data, as well as data from Mastermind and Genomenon, that we partnered with Genomenon to pull those data in. Those data also were a bit richer than what gets put into ClinVar, for reasons we spoke about earlier, we don’t get all the evidence necessarily being submitted. I do think that the scientists and clinicians that are maintaining ClinVar do a fantastic job. It’s just that, to have something that’s at that volume, and you have to accept what you receive, it’s a challenge. We do need to look at scalability and using informatics to pull in information more in the future.

LAUREN: All right, well, we have come up on the hour now, so that’s all the time that we have for questions today. Great discussion! I want to thank our panelists and our audience for participating today, and for your attention. If you liked today’s program and you’d like to participate in an upcoming Rare Perspectives roundtable, please reach out to me. I also welcome your ideas for future topics to debate. Before you leave, also, please take a moment to answer our brief survey. We appreciate your feedback. Thank you again and have a great rest of your day, everyone!

MARK: Thank you, Lauren. Thank you, Nicole. Thank you, Heather.

NICOLE: Thank you, have a great day. Enjoy Rare Disease Day tomorrow.