Our next speaker is Dr. Marty Tristani from University of Utah, and his topic is mosaics, how, how common, and how important. And you'll notice his slides are not his own. He's pinch-hitting for us. The original speaker had a family emergency. So thank you, Marty, for doing this. So this takes a little while to go through that disclosure there, but in the meantime I'd just like to follow up on, on Al George's talk and this concept of VUS, and, and kind of Greg, what you brought up as well, people are sitting on a lot of VUSs, and the key thing is, is how strongly do you believe that phenotype matches the genotype? A patient with, you know, has CPVT and has a RYR2 VUS, that's likely, you know, based on additional pieces of information, you could potentially make a call that that's likely pathogenic based on the phenotype that you're seeing, and the genetic companies will never do that because they're not seeing the patient. So and when thinking moving forward, we have to have a method, sort of a collaborative method for adjudicating everything together, you know, and sharing that information across different centers. So that's my spiel for, for VUS analysis. So, so I was asked to step in and, and talk about mosaicism, and so this is Facilios's slides, and I'll just talk about mosaicism and what is mosaicism. Whoops, sorry about that. Okay, so what I thought I would start off with is, you know, before we talk about mosaicism, which is a complex topic, just talk about something which is a little bit more simple, and that's just a de novo mutation. So here we have two parents that are phenotypically normal, and then we have a child that has a particular disease, and when we sequence them, they have a, this individual has a mutation which the parents don't have, so this happens spontaneously in the child. So what's the difference then with mosaicism? So mosaicism is when you have multiple populations of cells that have different DNA sequences. So in this example here, we have the father, and the father has some tissue, some cells which express a particular mutation, and the other cells are wild type. And if any of those mutations land in the gametes, then that person, that father then can pass on to the child a gamete that carries the mutation, and then the individual down here will have the mutation in all the cells of the body. So that's the germline mosaicism. So what about somatic mosaicism? So in the example over here, we have a situation, again, both parents are normal, and in the child there was an event that occurred that resulted in a mutation being introduced during cell division that then results in specific cells in the body having a mutation or having the wild type. And in thinking about mechanisms of mosaicism, these can be aneuploidy, this is an example of absence of the X chromosome, and that can give a phenotype which alters the pigmentation of the skin. So in this case here, what you would see is certain skin cells which have normal pigment, and certain skin cells which have abnormal pigmentation. And that's a very classic, easy, visual evidence for mosaicism, but that's actually pretty rare. The more common, well, common is still rare. The more common form of mosaicism, though, is a post-zygotic de novo mutation. So in this case here, we've got the gametes which are normal. They fertilize an egg. We've got the zygote. And then a de novo mutation occurs in one of the early cell divisions. And then that goes on to populate multiple tissues. In this example here, it would be the heart and the brain. And if the de novo mutation happened later in cell division, and these cells then are just committed to a cardiac lineage, for instance, then you would have mosaicism in the heart only. So this is heart-only mosaicism. This is multiple tissues. And in this case, you could have, you know, heart, blood, brain, skin, multiple areas. So how are mosaic mutations detected? Well, that's just with a standard sequencing methodology. And the term variant allele fraction refers to the percentage of reads which carry a particular nucleotide. So if a person is homozygous reference, so they have two copies, each of them have, say, from this example here, let's just say a T, then their variant allele fraction is one. If they're a heterozygous carrier for a particular genetic variant, then it's 0.5. And so this would be an example here, where they're a homozygous reference. And down here, what you can see is a little tiny signal on the chromatograph here. About 5% of all the reads localized to this here. And that's suggestive of mosaicism. This is difficult to determine, though. It's easy to overlook something which is this small. And so I think that's one of the major challenges when we look at sequences or when we get a genetic report back, that, for example, a variant is de novo, because it's not present in the parents. Is it possible that there's a tiny little signal down here that was missed? And that would suggest that the parent is mosaic and would have implications for how you're going to counsel the other children. So how can you detect mosaic mutations? Well, this is an example which would be, for example, say a tumor. And in this case, you've got lots of cells. You can sequence all the DNA, sequence all the cells. Or you can disperse the cells, grow them up, and sequence them individually. And the problem is that's very labor intensive. It's very expensive. And it's hard to do outside of the cancer world. There's lots of pitfalls in mosaic detection when you use those kind of methodologies. If you're passaging cells in culture, you can introduce mutations. There's a lot of false positives. And so the whole field of how we detect mosaic mutations is complex. This is an example of a patient that has Lanky T syndrome. And this is a paper that was published by James Priest at Stanford. They had a patient who was born with 2 to 1 AV block and very long QTC interval. They sent off a commercial Lanky T panel, but they also had access to rapid genome sequencing. So they did both simultaneously. So when the results came back, both of these methods actually detected a relatively common polymorphism in the SCN5A sodium channel gene. But the WGS also suggested an additional variant which was not flagged by the genetic counselor report. And that's really what's shown here in this slide here. So the whole genome reads in this case here is just saying what each read has at that particular locus. And in this case, there were 26 Gs and there were 7 Ts. So if this was a heterozygous mutation, you would have close to a 50-50 split. Not quite, but close. But this is quite imbalanced. And 7 reads for whole genome sequencing is, you know, you could say, oh, that's just an error in the sequencing. And you can see those errors, but it's enough to want you to take a closer look. And indeed, that's exactly what they did. So they went on then to take various samples from blood, urine, and saliva that correspond to different germ lines, mesoderm, endoderm, and ectoderm. And then they looked carefully at the percentage of reads which were the mutation. So here's the parent. Here's the father's saliva, mother's saliva. That's all the wall-type read. And then you can see varying amounts of DNA reads from the mutation from different blood draws, saliva, and from urine, again, suggesting that this patient is mosaic. So what was also unusual about this patient is the patient went on to develop a dilated cardiomyopathy. And in the paper, that's not fully explained why they thought that was and whether it was related to Long QT syndrome or did the patient have something else. But be that as it may, they actually had tissue, the heart tissue, that they could measure the reads of the mutant. In this case, it's a valine 1762 leucine. And they found that in the left ventricle, this is what, maybe 20% of the reads are mutant. And in the right ventricle, it was a little bit lower. So it brings up an interesting question. So how much abnormal gene expression do you need to have a clinical phenotype as dramatic as this kid had, which was 2 to 1 AB block? And I actually don't know the answer to that question myself. They did go on to functionally characterize by measuring currents. And in this case, remember, the patient actually had this common polymorphism. And so they measured the electrical effect of that, which was essentially normal. But the mutation had this standing current here. So this is the gain of function that we would expect with Long QT syndrome type 3. But the main question, I think, remains unanswered. And that is, getting back to this, what would this really predict about the heart? And how much abnormal transcript did the patient have? And was it enough to give him that degree of QTC prolongation? Moving on with another point about mosaicism, this is a paper that was looking at postnatal knockdown of junk to fill in in cardiomyocytes in a mouse model. And those mice go on to develop heart failure and die within 15 days. And moving into more of the mosaic realm, another group basically did a similar approach. But in this case, what they did is they used an adenoviral construct to infect the myocardium with CRISPR-Cas9. So then the myocytes containing the CRISPR-Cas9 and the guide will cut at specific regions in the junk to fill in to a gene and result in basically a mosaic. And their point here was to try and find out what was the threshold for which the degree of DNA damage that you had would then correlate with the phenotype, in this case, postnatal death. And so they were basically able to show that with low expression of the mutation and mid-expression of the mutation, those are sort of overlapping right here with 100% survival out to 28 days. But as you get higher and higher into the degree of deletion, then you get more and more death. And so this high group right here that had would be right here showing lots of the cells actually had the deletion did impact the survival, but not as much as when you get the full knockout. So the idea is that would be potentially a mechanism of how you could describe what would be the threshold if you had a patient that was mosaic for a particular disorder. So here's an example. And the question is, is this a de novo mutation or is this mosaicism? And here's the family here. We have one individual right here who died at 10 years of age while running. And we have another individual here who's now the proband who has recurrent syncope with an exercise treadmill test. And the concern is for CPVT. So even with this kind of model, you would probably guess right offhand that this wouldn't be de novo because you've got another affected individual in the family. But assuming you didn't have that, you might think it could be de novo. But I think this case would suggest this is some kind of inheritance pattern. And this is what the DNA sequencing looked like. So in the affected individual here, we see two peaks of equal height. So there's heterozygous for that mutation there. And then this is one of the parents here, right there. And you can see this very low expression or this very low signal from the mutation, again, indicating that the parent is mosaic. And then in that individual, they went on to look at different tissues, buccal, urinary, and leukocyte DNA. And it doesn't show up very well. But all of these have a signal down here which shows, using high-resolution melting curve, that there is some abnormal transcript there in the parent. So germline mosaicism. So how common is mosaicism? Looking at the autism data, it suggests that almost 8% of de novo individuals or de novo mutations in autism patients were postzygotic mosaics. In an analysis for long QT syndromes, more relevant to us here in the room, it was rare for individuals. So that's 0.05%. So the question is, is this a—are we missing this? And that's a possibility if we're not carefully looking at the DNA sequences from the parents. So finally, in conclusion, the question is, the majority of ryanodine 2 mutations are thought to be de novo. But could it be possible that some of these are actually germline mosaicism? And I think that's definitely important to know as far as counseling the family as to the likelihood of having another child with that disorder. And then the other question is, getting back to the other 25% to 30% of gene-negative long QT syndrome, we heard earlier this morning that perhaps there could be a contribution of multi-oligogenic contribution. And then the other possibility is, could there be somatic mosaicism or a mutation which affects the heart cells alone but is not present in the blood cells, for instance? So I'm going to stop there. Thank you. Marty, that was great. Thank you for chipping in. It was like they were your own slides. Any questions or comments from our panelists? Pete? Yeah. Marty, thank you for accelerating my timeline for retirement. I'm still trying to find out if Mendel's peas are bluer than yellow and why. With the mosaicism, particularly like in the long QT, rather than being how much is enough, in your opinion, do you think it could also be sort of a bell-shaped curve where you sort of are starting to increase the disparity and repolarization if you sort of have patchy effect in muscle, and it may be that there's a sweet spot where if you have a certain amount, that's the real danger zone. That is absolutely correct. Yeah. And I think that's a great explanation for James Priest's data, that you have cells which carry the mutation that are actually populating the key region of the heart or a particular region of the heart which adds to the arrhythmogenicity of it. So, I'm sorry. In what arrhythmia diseases has it been described, mosaicism, as actually causing disease? So we had a report in Anderson-Tewell syndrome. There was the child presented with severe long QT syndrome. Timothy. Timothy. I'm sorry. Timothy syndrome. Severe long QT syndrome, syndactyly, and 2 to 1 AV block. And the dad had syndactyly as well, and he was otherwise completely normal. And we looked in, we did blood, saliva, and sperm. And he was mosaic, and it was in the gametes. So I think there's probably multiple examples of diseases across the spectrum that we deal with. Do you think it could happen in any form of long QT syndrome, any form of inherited arrhythmic disease? Sure. Absolutely. His EKG was normal in the dad? No, he's borderline. The dad was borderline. Yes. Yeah. Not the strong phenotype. Correct. Correct. And he didn't have the intellectual disability, although he was sort of a Dungeons and Dragons guy, for what that means. Does that qualify, then? Are you a Dungeons and Dragons guy? Sorry. No offense to the Dungeons and Dragons people. Go ahead. Does Game of Thrones qualify? Here you go. All right. All right. Dragon fascination. Just to say that in RYR2 CPVT, there's, there seem to be a few lurking there, and there may be some data soon for the Royal Brompton Lab from that point of view. And we've seen a little bit in the sudden death group as well. And maybe those are things that we may be missing, and next-gen sequencing is going to help us find those, particularly when you find that so much of the serious disease is unheralded in the rest of the family. But you may see some minor manifestations when you look very closely. Yes. Yeah. I think it's very challenging. I know when I'm going through, you know, our data for like WGS, and you say, oh, is that real? Oh, no. The allelic balance is so off. It's probably just an artifact. But maybe not. So we have to be careful. Okay. Thanks, Mark. Okay. Thank you.

mosaicism

germline mosaicism

somatic mosaicism

mutation

sequencing methodologies