Welcome to the ACMG Genomics Case Conference. This is a monthly series planned by the ACMG Education Committee. Each month, the conference rotates amongst laboratories that perform genome-scale testing and the clinicians involved in the cases. Today's event is for educational purposes only. My name's Elise Mitchell. I'm a genetic counselor in the Genomics Laboratory at Mayo Clinic. My co-presenter today is Dr. Nicole Hottman, a co-director of the Genomics Laboratory at Mayo Clinic. Dr. Hottman and I have nothing to disclose. These are the disclosures of the planning committee. Today's conference is approved for one CME, and for those who applied for the credit, you will receive an email after the event with an evaluation and questions. These are the overall learning objectives for the case conference in general. And today's specific learning objectives are to discuss the gaps that exist in cytogenetic testing, to describe mate pair sequencing and how it works and might differ from other types of whole-genome sequencing, and identify situations in where mate pair sequencing might be of benefit. If throughout today's lecture you have a question, there is a Q&A box in the top right corner of your screen. Feel free to enter your questions there, and we will answer them verbally at the end of the presentation today. So now for the clinical utility of mate pair sequencing. This first image here will be familiar, if not to most people, to everybody on the call today. This is a G-banded karyotype, and this is, of course, the gold standard of cytogenetics that we've been doing for decades. It was actually the first genome-wide test. And with a karyotype, we can see large abnormalities like aneuploidy, large copy number changes, and we can also see structural rearrangements, whether they're balanced or unbalanced. The limitation of the karyotype, of course, is its very low resolution. We're able to see abnormalities that are greater than 5 to 10 megabases only. The second tool for cytogenetics to come along is FISH, and FISH testing allowed us a much better resolution. We could see deletions or duplications less than 200 kb, but the limitation of FISH is that you have to have a region to target that has a FISH probe available. Genomic microarray is kind of the best of both of those worlds. It allows us to evaluate the entire genome at once. It's ultra-high resolution. We're able to see copy number gains and losses of less than 10 kb. The limitation of a microarray is it's not able to detect balanced rearrangements. In regard to balanced rearrangements, we're going to be talking about them today throughout the event. Majority of the time when we find a balanced chromosome rearrangement, it's in a phenotypically normal individual. They're usually having karyotyping due to recurrent pregnancy loss or infertility or possibly a follow-up if they've had an offspring with an unbalanced rearrangement. They're known to be phenotypically normal individuals, so when we see a rearrangement and it appears balanced, we assume that it truly is a balanced rearrangement. On occasion, we find these rearrangements in a phenotypically abnormal individual, and then it's unknown whether or not that rearrangement might cause or be contributing to their phenotype. The issue is the chromosome studies can find the rearrangement, but it can't identify what genes are at or near the breakpoints of the rearrangement. Some rearrangements, some exceptions to that rule are in hematology when we do chromosome analysis for hematologic conditions. There are some very classic rearrangements that we see routinely and usually with the same diagnosis, so for example, a 922 translocation in CML. In those scenarios, you can infer what the genes are involved in the rearrangement, but that generally is not the case in a clinical situation or, sorry, a congenital situation, and it leads the families to a diagnostic odyssey. An example of a case that could benefit from MATE-PAIR, for case number one, this is a prenatal case. It's an 18-week fetus who on ultrasound had a structural brain malformation. There were no previous family history reported of congenital anomalies in the family. A karyotype was done and the fetus was found to have an apparently balanced rearrangement, a translocation involving 1 and 9 at the breakpoints listed in the nomenclature. Tests were tested and it was found that the translocation is de novo. Next step was in a chromosomal microarray to look for any deletions or duplications at that breakpoint or anywhere else in the genome that might have explained the brain malformation, and those results were normal. So while you're able to possibly provide the family with an empiric risk, it's very empiric and doesn't take into account anything specific like the brain malformation that they're seeing, and it would lead to really a diagnostic dead end in this prenatal situation. So I'm going to turn things over to Dr. Hottman to talk about MATE-PAIR sequencing. Thanks, Elise. So just to give a little bit of background on MATE-PAIR sequencing, MATE-PAIR is a specialized library preparation technique followed by whole genome paired and next generation sequencing. The difference for MATE-PAIR sequencing and what makes it work for these types of purposes is that the input DNA is quite large on the order of three to five kilobases as opposed to the 200 to 500 base pair fragments that are used for most standard NGS applications such as vanilla paired end sequencing. So this enlarged input size is what allows for the efficient detection of structural variation in the genome at a much lower sequencing depth than other NGS methods. So just to walk through the library preparation technique and how it works, here is a visual. So we start with an approximately three kilobase fragment, and I'm showing a red and a blue bar at the end of them to just signify pieces of genomic DNA that are far apart in the genome. These fragments are labeled with biotin, and that fragment is then circularized to bring the two ends together. These circular fragments are then fragmented again, and a biotin capture is performed so that we are left with three to 500 base pair fragments of DNA that are actually comprised of pieces of DNA that were originally three kilobases or more apart in the human genome. So we take those fragments, and then we put them on an Illumina sequencer or any sequencer that you would like, and we do paired end sequencing. And so we're sequencing 100 base pairs on either end, and we take those reads, and we map them to the human genome. And so what we should see is that those reads should map back to the same chromosome approximately three kilobases or so apart in a very specific orientation, so that the first read is in the reverse direction, and the second read is in the forward direction. This happens to more than 80% of our reads that we sequence, and this is what we call a concordant fragment. This is what we expect to happen most of the time. However, we also have what we call discordant fragments, and this is when paired reads either map to different chromosomes, such as the example I'm showing here where one end of a piece of DNA maps to chromosome three and the other is mapping to chromosome seven, or they could map in different orientations. So both reads are mapping in a reverse direction, for example, but still in the same chromosome. So this would tell us that we have a potential translocation with a breakpoint in between these reads or an inversion if they're mapping to the same chromosome, or et cetera. For example, we can also see copy number changes in this manner. So it's possible to tease out by looking at the mapping patterns and the genomic location what the rearrangement is and what the genes are affected, and we can usually distinguish a breakpoint within two to three kilobases or less. So I'm not going to go into too much detail about our algorithm, but this was a mail-built algorithm. There are other algorithms out there that are available, but we use one called binary indexing mapping algorithm, and I'm showing the publication here for anybody who might be interested later in looking this up, but I'm going to actually move past this and talk a little bit more about what we call our junction plots because we're going to be showing several of these throughout the presentation. So a junction plot is what we use to visualize where the breakpoint is in the genome or on the derivative chromosomes. So on the top right of the slide, I'm showing that same example with a 3-7 translocation, and down on the bottom left, I'm showing the derivative chromosome 3 and chromosome 7. You can see a breakpoint on the P arm for both of these. What this indicates is that we have a derivative chromosome 3 with some chromosome 7 material at the top, and likewise, we have a derivative chromosome 7 with some chromosome 3 material at the top. When we do mate pair sequencing, this is a whole genome sequencing assay, so we're sequencing all of the derivative 3, all of the derivative 7, and all of the other chromosomes in the genome. However, some of these reads that are marked by the green rectangles are actually spanning that breakpoint, so that one end will map to one chromosome, and one end will map to the other chromosome. So on the derivative chromosome 3, the blue circles are mapping to chromosome 3, while the red circles are mapping to chromosome 7, and vice versa on the derivative chromosome 7. On the far right, you see how we visualize this in our software. So we have the lower chromosome number on the top, the higher chromosome number on the bottom, and the lines in between the red and the blue are showing that the red and the blue reads are found on the same piece of DNA. So by looking at where the red dots turn to blue on the top, and where the red dots turn to blue on the bottom, this is where your breakpoint is. And you can then see easily, for example, on chromosome 3, there is a gene that's disrupted by this rearrangement. So one more concept to talk about before we go back to our cases is the concept of bridge coverage. So this is why mate pair sequencing is A, so efficient at capturing breakpoints, and B, at a much lower cost. So I'm showing a breakpoint on a random derivative chromosome between A and B in this example. And on the top is just regular paired-end sequencing, and on the bottom is mate pair sequencing at the same depth. So you can see that since the fragments that you obtain by doing paired-end sequencing are significantly smaller than what we start with for mate pair sequencing, you're much more likely to capture a breakpoint with high confidence by using mate pair, because those fragments are much longer. So I'm going to go back to case number one. So case number one is that 1,9 translocation that Elise spoke about previously. Again, this was found in a prenatal sample, and no further testing was available. It was found to be de novo, and microarray was normal. So as follow-up information that we obtained, the infant was born with a structured brain abnormality and microcephaly. We were then able, on the back end, to perform mate pair sequencing to characterize that 1,9 translocation. And this is a whole genome view showing a translocation between chromosome 1, which is on the top, and chromosome 9, which is about halfway down your screen. You can see the line between them indicates rearrangement. And then when we look at our junction plot, we see chromosome 1 on the top and chromosome 9 on the bottom. And on the top, you see that the chromosome 1 breakpoint disrupts the gene AKT3, and this is an intron 1. And on the bottom, the breakpoint on chromosome 9 is actually intergenic, so it is not disrupting any gene. And this was also confirmed by Sanger sequencing. Interestingly, this haploinsufficiency of AKT3 causes a syndrome that's well-known, called Omegalencephaly Polymicrogeria Polydactyly Hydrocephalus Syndrome 2, or MPPH2. And the clinical features of this syndrome fit very well with the fetus in this case. So the take-home point here is that had this assay been available previously, this information could have been provided in the prenatal period when the de novo translocation was originally detected, and this would have allowed for much better-informed pregnancy decisions and probably a lot less downstream testing being performed. So I wanted to highlight one of the other situations in which mate pair is useful, because folks in the field actually come across these quite frequently. Mate pair is actually quite useful at determining orientation and genomic location of duplications. So I'm showing a duplication on this slide here, and you can see there's a copy number gain and a gene on the bottom, towards the bottom of your screen, that shows that this duplication is including part, but not all, of this gene, and it's not included within the gene. So when we see these in the clinical laboratory, we typically sign these out as uncertain, because we don't know where that genomic material is and what it's doing to the gene that it's overlapping. And I'm showing here a very nice figure from a paper that was published in the American Journal of Human Genetics by Newman et al., and this is a nice figure that I'm not going to walk through in totality, but it shows what the different types of duplications are and what effects they have on genes. So if we start on the left, you can see what a normal locus looks like, and then just to the right of that, you see that when you have a direct duplication, and this duplication overlaps a portion of the gene, in general, these duplications do not actually disrupt the genes. They can create what we call fusion genes, the clinical significance of which is uncertain, but in general, they don't have an effect, whereas if you move slightly to the right, you can see that when we have an inverted duplication, these often do have the potential to disrupt the genes that are located on either side of the duplication and can result in haploid insufficiency of an important gene. And then on the far right, we have the case of insertional translocations, but the take-home point is that when we see a duplication by microarray, we have no idea where that extra material is and what the orientation is to tell us whether or not it is inverted or tandem. So in most cases right now, we just reflect to parental studies to show whether or not it was inherited or de novo and hope that we get good phenotypic information on the parents to allow us to either bump up the likelihood of it causing disease to likely pathogenic or decrease down to a likely benign. But often parents are not available or we often don't have good phenotypic information on the parents. So Newman et al. did show that most duplications are tandem and direct, and this has been our experience here at Mayo as well, and do not disrupt genes. The remainder of the findings in their paper were either more complex, were insertional, or inverted that potentially would disrupt a gene, but again, we can't tell by microarray what the case is. So this is case number two, and this is a two-year-old male with developmental delay, speech delay, dysmorphism, and macrocephaly. So chromosomal microarray was ordered and performed, which found this duplication on the X chromosome, and it's difficult to see, but at the bottom you can see that this duplication partially overlaps the BRWD3 gene. So BRWD3 is important because haploinsufficiency of this gene has been reported in males with intellectual disability, macrocephaly, and other features. However, like I've said multiple times now, we had no way to determine whether this duplication disrupted BRWD3 gene function. So in this case, we would normally test the mother first, because since this is a male, it would have been inherited from her. And then usually what happens is that mother is found to carry the duplication, but then since X inactivation is a factor, we would then have to go on and test multiple other males in the family, hoping that somebody ended up being informative. In this case, it was maternally inherited, but again, there was no additional information that was derived from that. So we were able to run mate pair, and what we found is that this duplication is found to be a blue-to-red duplication. And so what that means is that you can see at the top on the right that this duplication involves genomic material downstream of BRWD3 and the 3 prime end. And just below that, I'm showing the duplicated material. And what happened was that this was inserted downstream of BRWD3. And you can see the red-to-blue where it's calling a discordant read. And you can see that although there is a duplication and it duplicates part of this gene, the original gene shown in blue on the right is completely intact. So we would interpret this as likely not disrupting BRWD3. And so we cannot say that this is pathogenic. So I wanted to touch a little bit on oncology, specifically a hematology case. I had anticipated that the audience today would be more interested in congenital rearrangements. But this technology is actually very useful in the case of balanced chromosome rearrangements in oncology and hematology as well. So this is an example of a 5-6 translocation. And we see these very frequently in neoplastic disorders. So when we see these in a cancer setting, when we see these balanced translocations, assuming that they are not inherited or congenital, these can indicate either chimeric fusion genes that are created, disruption of tumor suppressor genes, activation of oncogenes, or sometimes they're just passenger mutations that they happen to come along and they are part of the clone but not necessarily important. When we see these, if they are disrupting an important gene or they are creating a novel fusion, this is important because that information can impact diagnosis, prognosis, and or therapy in many cases. I'm going to talk about a third case now, which is a 10-year-old female. So this girl had a new case of BALL several years ago. And at the time, the test that was ordered were chromosomes as well as our B-cell ALL fish panel. Both of these were done on bone marrow. So I'm showing you the chromosome study here. And what we saw was a hyperdiploid clone with a few abnormalities, only one structural. Otherwise, it had multiple copies of 12 and chromosome 15 and markers as well as trisomy 8, but nothing really of note. The fish panel was normal except for ETB6. So what I'm showing here on the left is an interphase image of our break apart fish probe strategy for ETB6. And this is showing a disruption of the ETB6 gene. What we do when this happens is we usually then will do fish on a metaphase to see if we can identify what the partner gene of that might be. And so in this case, we did that. And you can see on the right, you can see some normal chromosomes 12 with a fusion signal indicating they're normal. And then you can actually see two abnormal chromosomes 15 with only one 5 prime ETB6 signal. So what that's telling us is that there's an unbalanced rearrangement and that there is some fusion partner on chromosome 15, but we're not quite sure what it is. So usually when this happens is that we'll go back to our fish probe arsenal and we'll say, is there a fish probe on chromosome 15 that can help us potentially identify what this partner is? And unfortunately, we didn't have any. So like I said, ETB6 was notable in this case because it has been disrupted in multiple tumor types, both solid as well as hematologic tumors such as ALL. ETB6 is an interesting gene because it has multiple partner genes. And it also has multiple mechanisms. So it may result in constitutive activation of an oncogene. It can result in loss of function, activation of nearby oncogenes, et cetera. But by itself, this rearrangement has uncertain prognostic and therapeutic significance for this patient. Right now, all we have for this patient is a way to monitor minimal residual disease. But it's not giving any indication as to what treatments might be useful for this patient, et cetera. So like I said, the partner identification could help, but we have absolutely no fish probes that can help us here. So this case was then sent back to us a few years later as she had relapsed. So a relapsed bone marrow specimen was sent to us for testing in 2017. And we ran both chromosomes in fish. And we noticed that fish studies were positive with the exact same pattern for ETB6 break apart, and no other abnormalities were observed. So we did mate pair on this case. And at that point, we were able to see, and this is the whole genome view at the bottom, we see a 12-15 translocation. And then if you look at the junction plot, again, chromosome 12 is on the top. And chromosome 15 is on the bottom. And we see very easily that chromosome 12, the break point is in ETB6, as we expected. And on chromosome 15, the break point is actually in a gene called NTRK3. So this would be predicted to result in a fusion gene, a chimeric gene consisting of exons 1 through 4 of ETB6, and exons 15 through 20 of NTRK3. And this is important because it includes the entire tyrosine kinase domain of NTRK3. So this would be thought to cause constitutive activation of a tyrosine kinase, much like the 922-BCR-ALL fusion does. So this particular fusion is extremely rare. It has been reported in patients with B-cell ALL as well as other neoplasms, such as high-grade gliomas and secretory breast carcinomas, as well as AML. The important thing here is that this fusion is actually targetable using tyrosine kinase inhibitors, such as crizotinib. So the take-home for this case is that this patient may have received different targeted therapy at the initial diagnosis had mate pair testing been available at that time. So I'm going to hand it back over to Elise to talk about the clinical experience that we've had to date. Thank you. So a little bit about our clinical experience up to now. When we validated the test, we had 75 specimens that had known rearrangements from either chromosome analysis or microarray. And we found a 96% reduction in the number and we found a 96% sensitivity, meaning we were able to validate and find the same finding on mate pair sequencing in 72 out of the 75 cases. The gap that we were not able to were either cases of low-level mosaicism or cases that involved a repetitive area of the genome. So Robertsonian translocations, for example, are not generally detected on mate pair screening due to the repetitive nature and a few other of the areas as well. As for the clinical cases that we have done in our laboratory, we've done 13 to date. 12 of them were done as a follow-up to an abnormal karyotype. And one was a follow-up to an abnormal microarray. Of those 13 cases, six of them we were able to classify as likely pathogenic or pathogenic. Seven of the 13 remained uncertain. And there were a few other interesting findings, too, that have come out of these 13 cases, which I'll give you some quick examples of. Three of the 13 were found for the abnormalities to be significantly more complex than what we had originally appreciated. And three of the 13 invoked the potential that it could be a positional effect causing the phenotype. For this case here, case number four was a newborn male with subtle dysmorphic features. On karyotyping, he had an inversion of the X chromosome on X with the exact nomenclature there in the report, or in the exact breakpoints in the nomenclature. We did mate pair sequencing and found that there was a gene disrupted with MAP7D3. But in the literature, there was only one patient described, and it was a putative loss of function mutation. So we described it as uncertain and did not feel that that necessarily we could say for sure if it explained his phenotype. However, in doing the mate pair, we actually saw that he had two inversions. One was the paracentric inversion that we appreciated from chromosomes, and then there was also a paracentric inversion that had not been noticed before. Case number five is an example of a possible positional effect. This actually started off as a hematologic study on a bone marrow. This was a 16-year-old female who presented with a headache, abdominal pain, vomiting, and had a possible diagnosis of porphyria. We did her bone marrow chromosome analysis and found an inversion. And around the same time, and that inversion was in every metaphase, and around the same time, her pathology testing found no evidence of hematologic disease. So taken together, we speculated that the inversion that we were seeing was, in fact, possibly a constitutional abnormality or congenital. Mate pair sequencing was done, and the genes at the break point were of unknown significance. But what was interesting, given the possible diagnosis of porphyria, was that we found that the ALAD gene in acute hepatic porphyria was located just nine kilobases from the break point, causing us to speculate that it could potentially be a positional effect. This is an autosomal recessive condition, so we recommended sequencing of the gene to see if there was another mutation in the other allele. On the screen now is our algorithm for when a balanced rearrangement is found in a patient with an abnormal phenotype. We do have another algorithm underway for when a duplication is found on microarray. That is in the works, but it's going to be pretty similar to this one. And really, the take-home message here is that the interpretation of a mate pair is at its best when the rearrangement is known to be de novo. So if both parents are available, we recommend karyotyping on both parents first. And if the rearrangement's de novo, then mate pair sequencing. If it's inherited, the question would be, is it inherited from a parent who is phenotypically normal or abnormal? And if the parent has the same abnormal phenotype, then again, mate pair would be recommended. If it's inherited from a parent with a normal phenotype, the next step might be array to see if there's any other gains or losses that could possibly explain the phenotype. If neither parent's available, you can consider doing microarray first or also doing mate pair. The mate pair interpretation might have some disclaimers if it's not known to be de novo, but still could be done. But it might also be reasonable to do a microarray first or instead of that. And for scenarios when one parent's available, you end up kind of in a hybrid algorithm, depending upon whether or not it's inherited from that one parent who's available. So in conclusion, mate pair sequencing is a useful tool. We can use it to reflex, to help clarify uncertain findings on other cytogenetic testing. Significant proportion of cases are more complex than what we had originally thought. There are still gaps that exist. So additional methods, like expression studies, might still be useful in some situations. I have a slide here with a few references and a slide with acknowledgments. There are many people here at Mayo Clinic who are instrumental in the development of this test, as well as the implementation of Go Live and are involved in the current clinical cases. We have many development techs, genetic counselors, laboratory directors, bioinformatics, and an entire other group. Hopefully, we've included everyone on this slide and our apologies to anybody who was accidentally missed. So if you do have any questions, we are happy to answer questions. You can type in a question on the Q&A section. There should be a Q&A box on the top corner of your screen. OK, this is Nicole again. The first question is, what has been your experience with insurance coverage for clinical mate pair sequencing in constitutional cases? That's a fantastic question. We have found that it's actually all over the board. We did apply for PLA code, which is a proprietary lab analysis code that can be used when only one institution is currently performing an assay or a technology. So to get a CPT code for a new test, you have to have multiple institutions who run it. We were able to get a PLA code instead. What we have found is that the insurance, despite having the PLA code, sometimes it pays for it, sometimes it doesn't. And then the amount that it's paying is extremely variable, anywhere from a couple hundred dollars to a couple thousand. So hopefully, over time, that will standardize. But we're still working through that process. OK, the next question is, what depth of sequencing coverage is required? Also a really good question. So we sequence these. It's a little bit different than paired-end sequencing, where you say, I'm going to sequence to 30x coverage. The actual base coverage we're getting from mate pair is actually only about 6x. So the actual sequencing depth is quite low. We think more in terms of that bridged coverage. And so we get about 60 to 80x bridged coverage. So we typically have very high confidence in the rearrangements we're finding. If you think in terms of how many patients can you pool on a flow cell or in a lane, we multiplex two patients at a time. So another question, in case five, did you do a PHA-stimulated blood test? Did you do a PHA-stimulated blood karyotype? And if positive or familial, is there evidence that symptoms such as hers exhibit later in life? So the answer is no, we did not actually do a PHA-stimulated karyotype for her example. And that was just because of convenience. We had a bone marrow sample here. And so we just did mate pair on that. So we don't really know if it was inherited or not. It is a recessive condition, the hepatic porphyria. So I'm not sure if there was any family history. But it can be a somewhat later onset. And the diagnosis was actually pretty questionable in this case. But I'm not sure if that was because she just never was able to present when she was having symptoms for the biochemical testing to be positive or not. But we did recommend sequencing of the other allele. And hopefully, that will be able to help them yield an answer. OK, I have another question. Is the advantage of mate pair simply due to the distance between reads? And how much extra work does it take to do mate pair? So the answer is yes. The reason that we can detect breakpoints so reliably at such low cost is because we're starting with those large fragments. And so that really allows us to do less sequencing but still detect these breakpoints. Whereas I had done some work a few years ago with paradigm sequencing. And you had to sequence quite deeply and at a much higher cost to detect rearrangements. And you're still not able to detect them quite as reliably, at least back then, because your reads are just not very far apart. With respect to the extra work, it is a different library preparation. And it is quite manual. And it's about a two-day process to perform the circularization and then shear those circles. But the rest of it downstream is really just simple paradigm sequencing. So once the library preparation piece is over, it's very similar to what most labs are used to doing. The next question is, what is the smallest duplication you can identify exact breakpoints for? This is another really good question that we don't have a definitive answer for. We've not actually validated with precision what size duplication we can detect at what confidence threshold. We just have anecdotal evidence that we know that we can detect duplications of a couple hundred kilobases and higher probably 95% of the time at least. When they get smaller than that, we're just not sure how reliably and how consistently we can detect those. But you can always do things such as changing your pipeline criteria and doing additional sequencing to help detect those. All right. I'm seeing a question about the ballpark cost of the test. I'm not probably supposed to say, but if you're interested, you can look at our Mayo Clinic Laboratories website, and it should be listed there. I'm seeing another question, if you could increase your insert size to 1,500 base pairs, would you get the same result with normal paired-end sequencing? That's actually a really good question, and we had tried to do that several years ago, and it did not actually work quite as well. But we – honestly, we didn't give it a very good go, because we already had a group here that was very invested in MatePair and had a lot of experience with it, and had really refined the process and the pipeline to do the analysis very well. So I would imagine, yes, if you could increase that insert size for regular old paired-end sequencing, that the – I'm not quite sure if it would perform quite as well as MatePair, but it would definitely perform better than a standard 2 to 500 base pair insert size with regular paired-end sequencing. One other question is, are you validated for saliva? And at this time, no, we are not. I'm seeing a question about TADs, so do you include interpretation of the TADs? So these TADs stand for topologically associated domains, and the answer is no, at this point we do not. But what we do do is that when we see a rearrangement, we always start at the break point for that rearrangement on both chromosomes, and we start off with the question of, okay, what gene is disrupted, and what do we know about that gene? And from then on, what we do is we walk out from that break point to about 500 KB on either side, and we take a more global look at all of the genes that are within about 500 KB or so of those break points, and then we start asking questions about what do we know about those. When we do that, you know, you're inducing more uncertainty, because once you're not at a break point, you don't really know what's happening to a gene and what is happening to its expression, but that's where the clinical correlation comes into play. So we've had three cases now, like Elise says, where we have had enough phenotypic evidence that made us extremely suspicious of a case where the break point is probably not in a gene, but it very well may be influencing the expression of that gene. Could you use this to size triplet repeats in an unbiased way? I don't think so at the depth of sequencing that we are looking at currently. What we are looking at here at Mayo, and I know a lot of other institutions are doing this, is looking at things like PacBio or the long-read sequencing. I think that will be a lot more accurate in determining size of triplet and other repeat expansions. Have you detected any small nucleotide variance by mate pair sequencing? The answer is actually no, because we're not sequencing at a depth that we can do that reliably. We have talked about just increasing our sequencing to be able to do this. So, for example, if we did take a mate pair library and we sequenced it to a 30x depth, then we could definitely do this, but then we are increasing our cost by 5x, which at that point you might as well just do an exome or a genome, because it would be cheaper. Another question about turnaround time. So, what is the general turnaround time for testing, and will you do on prenatal samples? We quote a turnaround time of 21 to 28 days approximately, and that's somewhat long, but what that includes is, A, the initial sequencing and analysis, which takes about a week, also factoring in that we only run this assay once a week, because we don't have the volume to do it more than that. Then on the back end, once we get our results, we do always do Sanger sequencing to confirm what we're finding, and that also helps in cases where we do need to refine our break points. This happens more frequently for heme cases than congenital, because sometimes for a heme case, you really need to know what intron is a break point in, in order to determine whether something is in-frame or out-of-frame, so we do that as well, and that adds another week or two, because we have to design primers and get our Sanger sequencing back, and then finally issue our report. But in cases where we don't have to do that, our turnaround time tends to be much lower, closer to two weeks. And currently, we actually are validated for prenatal samples, we just haven't pushed that test to be live yet, but we will have a different order for prenatals, hopefully very soon here, but we are validated for them. So you have a question about our analysis software, so what is the analysis software, and is it available for others to use? So I showed the paper that explains our algorithm, which is on PubMed, and would be available if anybody actually wants a copy, feel free to email me. The visualization software was built here at Mayo as well, and there is, we have another publication that describes that. It's not currently available to anybody outside of Mayo Clinic, but I think there are some other visualization tools that other groups have come up with as well, I'm not sure about availability of those. A really good question about FFPE. Have you attempted any mate pair sequencing on FFPE specimens? The answer is yes, we have, and no, it does not work very well in our hands right now. And this is very unfortunate. We are working actively on developing a protocol for FFPE. It probably won't be mate pair, but it'll be something similar, for example, somebody mentioned paired end and hopefully increasing your insert size with FFPE, we couldn't increase it too much, too terribly much, but we are working on an alternate protocol to do that. One thing we are validated to do though, and we've had pretty good success with it, is actually fixed cell pellets. And so those of you who work in cytogenetic labs, you may know that often the only specimen you're left with after you do your chromosomes and fish is a fixed cell pellet. We found very frequently and very early on that we were getting requests to do additional mate pair testing on samples, and that's all we had. So we did validate that, and we've gotten it to work pretty well. It's a little bit noisier, but in this test, we know what we're looking for, and usually we are able to get an answer in that situation, so we are validated for that. And actually, I just saw a question about that, so that's answered. One question, is there a PCR amplification step in library preparation? And actually, yes, there is. The second part of that is, would it not affect characterizing GC-rich regions? And there is definitely some noise and some bias in the GC-rich regions. The PCR amplification we do is not an extensive PCR, so it's not like a 30-cycle PCR. And so, we don't notice much bias, but we do normalize to a group of normal individuals, and that helps get rid of those issues with the GC-rich regions. One question about selection of 3 to 5 KB fragments, so the question is how do you select for 3 to 5 KB fragments when building the library? So what we do is we have adjusted the shearing of those circles to give us 3 to 5 KB fragments. So if you shear too much, your fragments will be smaller, and if you don't shear enough, your fragments will be larger. So and there's definitely an art to that. We use a Kovaris, and you definitely have to do a lot of optimization to figure out what your shearing time is, because it's really easy to get those smaller fragments, and if you shear your circularization product too much, you'll get 1 to 2 KB fragments, and those 1 to 2 KB fragments just are not as good at capturing your break points, and so you have to do more sequencing when your library is that small to detect your break points confidently. Okay, looking for just a few more questions I might have missed. I'm seeing a few more questions on cost and someone was not able to find it on our website. I can't speak to Illumina's cost to do the testing but it's on the magnitude of a few to several thousand dollars. Okay, I see a question about testing mate pair on frozen blood specimens. We have actually not done this, but I would imagine it would work pretty similarly to that of a fixed cell pellet. It really just depends on the quality of the DNA coming out of that frozen blood specimen. All of ours have been received as far as I know ambient. We would try the extraction if somebody did send us frozen blood, and we would just look at the quality of the library we got out on the back end and see if it was suitable for mate pair or not. That's all the questions I'm seeing, so...one second. All right, so we're going to move on. I would just like to bring folks' attention to the 2019 ACMG Annual Clinical Genetics Meeting that's being held the first week of April, and that will be in Seattle. And I'd also like to mention our 2019 Genetics and Genomics Review Course that will be held June 27th through July 1st, and you can get more information on those at acmgmeeting.net. I just want to mention that a recording of this seminar will be available in the ACMG Genetics Academy. And at this point, I'd like to thank the audience for listening, and everybody, please have a great day. And like I said, if anybody has additional questions or a follow-up, I'd be happy to hear from you via email. So, thanks a lot. Thank you very much. We hope you've enjoyed the seminar today by ACMG Genetics Academy. Can you guys hear me? Hello?

ACMG Genomics Case Conference

mate pair sequencing

clinical utility

whole-genome sequencing

structural variations

gene disruptions

prenatal case

balanced rearrangement

fusion gene