Hello everyone. My name is John Bergerstein. I'm here with Matt Wheeler. We're at Stanford University at our Center for Undiagnosed Diseases. We're happy to be here today to present some of the work we've been doing as part of the clinical site for the NIH Undiagnosed Diseases Network. Before we start with our presentation, we've been asked by the college to share a few announcements and some background. So, a reminder that the abstracts for the next spring's meeting are due on December 8th. Today, I will be focusing on a series of cases that highlight steps that can be helpful in uncovering diagnoses, starting with a non-diagnostic clinical exome. Our talk today and the series has been sponsored by QIAGEN and the Injunuity Clinical Decision Support Platform. We've been asked to mention that today's presentation is meant to be informational. It's not meant to be relied upon to make clinical decisions and the cases presented here don't necessarily correspond to what others in our audience may be seeing back at home in their own clinics. Again, my name is John and Matt, Dr. Wheeler, sitting next to me. Both. Neither of us have any disclosures today related to our presentation, and we're appreciative of the planning committee and staff who supported getting ready for today as well as the other events in this series. Today's presentation will be one hour in duration and those who have signed up are eligible to receive one hour of CME credit. The instructions on how to obtain credit should have been sent to you at the time that you are registered and you should receive an email within 24 hours of the presentation in order to document that you've attended today. There will also be evaluations which you're encouraged to complete. So our series has a number of learning objectives and we'll be focusing on the first three today, focusing on the utility of exome versus genome sequencing, thinking about the indications that have the highest yield with respect to exome and genome sequencing, and some factors about assessing pathogenicity of variants. We're hopeful that at the end of today's session our participants will have a fuller understanding of classes of variants that are more likely to be overlooked or misclassified during routine analysis, describe findings suggested of the presence of mosaicism and some strategies that can be helpful in uncovering that mosaicism, and we'll be highlighting the strengths and limitations of various currently available diagnostic methodologies for uncovering different classes of genetic variants. At the end of our presentation, in about 40-45 minutes, we'll have time for questions which will be submitted through our web platform and you can enter those and we'll answer as many as we have time for. So we're speaking today from the Stanford Clinical Site of the Undiagnosed Diseases Network. We're one of seven clinical sites and they're distributed across the country meant to enable maximum access of patients and clinicians to the services available through the network designed to be helpful in uncovering the etiologies of rare and under or undiagnosed conditions. We're talking about the work of the clinical site, ours, and also how it's been supported by some of the core facilities of the network, which include sequencing facilities for exome and whole genome sequencing, as well as a metabolomics core facility, a model organism core facility, and a bio repository. The research network is also supported by a coordinating center at the Harvard School of Medicine. Now broadly, the goals of the UDN are really to improve the level of care and the frequency of obtaining a definitive diagnosis for patients, facilitate areas into the science of diagnostics, and also to develop a functional and durable model for collaboration around the diagnosis of rare diseases and developing therapies for them. So we'll be focusing today, as promised, on a series of cases that are illustrative of some of the points that we hope to get across in our center and across the network. We see patients of all ages, from infancy to adulthood, with indications for referral spanning the spectrum of medical practice. It's important to emphasize this is just a part of the network and benefiting from the strength of collaborations across the multi-institutional project. We are also working with a team of individuals from many backgrounds and specialties here at Stanford, including the full spectrum of clinical providers, a talented team of clinical study coordinators, genetic counselors, bioinformaticians. It's really a team effort. So the process by which patients may participate in the UDN begins with submitting an application that's submitted through the coordinating center. Applications are by default routed to the center closest to where they live, so it's geographic, unless there's a clinical indication that suggests that they may benefit from being evaluated at one site or another. Applications are reviewed locally, and a preliminary decision is made to accept to the network or defer the application. Subsequently, the cases in line to be accepted are presented on a national conference call from participants from all of the participating sites, where information is gathered on suggestions for possible diagnosis for a workup. The actual clinical and laboratory evaluation can occur in the span of a one-week visit or a series of ambulatory visits. In some cases, sequencing may be done first. In some places, a clinical evaluation may be undertaken in advance of sequencing, depending on the prior workup and the particular clinical situation of the patient. For patients' ambitions of undergoing clinical and molecular evaluations, typically blood and urine samples are banked. In many cases, fibroblasts and lines are also established. And on an as-needed basis, studies may be taken in model organisms, and it's important to emphasize that a foundational element of the network is the systematic sharing of data both within the network and in a de-identified manner. Data are also pushed periodically to dbGaP to be available to the wider research community. As of a month or so ago, this was the pattern of enrollment from across the country. Outreach efforts are going on. Referrals are welcome from across the entire 50 states, recognizing that some areas are already served well by facilities and resources outside of our network. There have been coming up on 2,000 applications, several hundred of which are currently under review. Nearly 500 evaluations have been completed, and over 100 participants have been diagnosed. Many of these individuals are people who have had exome sequencing and DNA microarray as well as other workups prior to entering the study. We're going to return now to the five cases for which a molecular diagnosis was not readily evident from whole exome sequencing data, DNA microarray, or other clinical studies, and look at how, ultimately, a diagnosis was arrived at. The first case is a boy who was presented at age 11. He had extensive venous malformations, including hypoplasia of the central nervous system. He had absence of the inferior vena cava as well as the right axillary vein. His skin was notable for the findings of hyperpigmentation of the skin, hypermelanosis, more specifically, which is evident on his right neck especially. He had nevus opulta, or hypermelanosis of the globe of the eye as well. He also had a nevus lamius, or capillary malformation, that was extensive over his right arm, trunk, and legs, and also on his left foot. We have some pictures to show you. The distribution of these findings was highly suggestive of a segmental or mosaic disorder. He had hypermelanosis of the eye as well as his right hemipalate. If we go to the next slide, we can see the extent of the nevus lamius on his right, and also more so, sorry, more on the right, and also on his left foot. We also have some evidence of hemiplegia of the right leg. On the left is a MR venogram showing that because he had absence of inferior vena cava, there were extensive superficial venous collaterals that were visible on his back. Also, on the right of the slide in front of you now shows the portal that's used for systematic collection of data related to patients evaluated through the EDM based on a phenotype system developed at the University of Toronto, and it enables to capture gastrometric information problem-less using human phenotype oncology. There's also the opportunity to display variants that have been prioritized or de-prioritized, and these data are accessible within the network and also are pushed to dbGaP. Ultimately, the EDM and all diagnostic enterprises are exercised in matching, and we're trying to systematize how we do that to facilitate it being done manually, and also ultimately to facilitate it being done in a manner assisted by computational tools. So, as this patient was evaluated with his distinctive features, we're able to be pulled into a picture consistent with the diagnosis of phacomatosis pigmento-vascularis, a condition that hadn't been described on phenotypic grounds for several decades, and I think opposed could be subdivided further based on the specific character of the vascular and pigmentary differences. The cardinal findings of phacomatosis pigmento-vascularis are hyposigmentation or hypermeiosis, necrosplamia, sometimes hemi-hyperplasia, absence of these veins, and in particular, the synovus of OTA of the eye. A classification system, a contemporary classification system, was proposed by Dr. Happel in 2005. So this patient's exome study initially was non-diagnostic. However, a deep sequencing study was done utilizing samples from a skin biopsy. A low level of those, approximately 5% of reads, was present in the affected tissue and absent in unaffected tissue. It was possible to validate this finding by a finger sequencing after restriction digestion of the wild-type allele. And the mutation was in the gene GNAQ, which concurrently with the evaluation of this patient was reported to be associated with both ester's rubella as well as the sectional disorders recognized as phacomatosis pigmento-vascularis. The non-diagnostic exome in blood, diagnostic in the affected tissue due to the mosaic nature of the patient's disorder. The next case we'd like to share goes in a different direction from a non-diagnostic exome. This is the case of a 7-year-old girl who initially presented very early in life at 2 days of age with a picture of a metabolic disease. She had hyperammonemia, lactic acidosis, briefly hypoglycemia, and over the course of many years had a number of biochemical enzymatic studies, molecular evaluation, including a microarray and a clinical exome sequencing and subsequent reanalysis of that data on a research basis were initially non-diagnostic. Following the patient's enrollment in the UDN, additional attention was paid to the possibility that there could be an element of consanguinity in the patient's family, and increased attention was paid to the variants that appeared to be identical by descent. In particular, attention was focused on a variant in the ATP5D gene not previously associated with human disease. In parallel, the patient had undergone a biochemical analysis that showed evidence of 3-methylglutakonic aciduria in a clinical urine organic acid study. Metabolomic profiling done locally as well as through the UDN core laboratory showed elevation of tricyclic acid cyclointermediates also suggestive of a defect in the respiratory chain. So we had a candidate gene with two variants in trans and supported biochemical evidence with a picture of a mitochondrial disease of the variant and a subunit of complex 5. However, we did not at that time have a second case. So case matching is a very important aspect of delineation and discovery of rare diseases, and that's through the UDN. This is attempted and supported in a number of ways. So we've talked about the systematic collection of data, and it's pushing both to the dbGaP. It's also pushed to the matchmaker exchange through Phenome Central. Additionally, patients and families have the opportunity to participate in a patient participant webpages project, which presents brief clinical synopses of their pertinent clinical history, which has succeeded in a number of instances in identifying case matches. So through multiple efforts, a second patient was identified in the U.K. who also had two variants in trans and ATP5E, completely overlapping spectrum of clinical findings. So we went on to pursue functional studies to additionally support the relationship between ATP5D and potentially a novel mitochondrial disorder. The model organism team of the UDN, led by Michael Wenger and Hugo Billen, developed a Drosophila model, which reduced the expression of ATP5D in trans patients. In a regional manner, shown on the upper left, the wild-type flies, and next to that the fly lacking ATP5D, demonstrating a very pronounced small head and small eye phenotype, which is shown on the upper right, could be rescued by the wild-type human ATP5D transcript, as well as by the transcripts corresponding to the mutation in the UDN patient and the patient in the U.K. So the head growth was restored, and the eye phenotype was incompletely rescued by the mutant transcripts, but fully rescued by the wild-type transcripts. So providing functional evidence in addition to the segregation evidence supporting the pathogenicity of the ATP5D variants. A third case, going along a similar direction, going forward from a non-diagnostic experiment, Exome, is the story of two siblings who initially had a period of typical development followed by developmental regression, with evidence of seizures and occasional myoclonus. In the older sister and the younger brother, the history was a little slower in that at the age of four there hadn't been evidence of developmental regression, but did have autism, intellectual disability, and epilepsy. Their evaluation had been very thorough, including biochemical studies, DNA microarray, a number of epilepsy panels, and eventually whole exome sequencing. There were a number of unaffected siblings in the family as well. This is a slide just to say that they had quite an extensive workup that has to be in very small print to fit on a slide. The patient underwent clinical whole exome sequencing, which re-identified a variant that had been seen on the epilepsy panels, a gene KCTD7 associated with a progressive myoclonic epilepsy, but a second variant had not been flagged. Re-analysis of the data following the UVN enrollment identified a second variant that was predictably synonymous in the same gene. This was a gene that was in ClinVar and had been reported as, sorry, a variant in ClinVar that was reported as likely benign. It was very rare. It was not present in databases. The informatics approach was predicted that it could potentially affect splicing, which made us increasingly interested in it, in addition to the phenotypic overlap between the patients and patients reported with bilayer amputations in KCTD7. The variant also segregated in the family, being present in trans the first variant and the two affected siblings, but not in the unaffected siblings. RNA sequencing evaluation was performed. Additionally, the transcripts were looked at by gel electrophoresis after RT-PCR, both of which showed that individuals who carried the CD156G to A variant demonstrated a typical transcript where there was exposure of a cryptic splice stripe upstream of the usual position, resulting in an abnormal transcript. Summarizing our first three cases, all started with non-diagnostic exomes. One proceeded to have a mosaic condition not present in blood. The second had a missense variant in a gene not previously associated with a disease that was exposed on a manual curation. The third had similarly a variant that had been overlooked on the basis of being predicted to be a synonymous of variant that was indeed a cryptic splice site donor. So the fourth case is an adult case that was referred. He's a 46-year-old male with a history of chronic hypokalemia, I'll use for the typo, metabolic alkalosis, hypocalcemia, and pseudo-hypoparathyroidism. He's had multiple non-traumatic fractures and over the course of the past 15 years has developed severe bronchiectasis. So as is typical of many adult patients we see, he'd had prior genetic testing focusing on the hypokalemia he's had. And he had genetic testing in five genes known to be associated with Barger-Gittleman syndrome, which was kind of revealing. In terms of the pseudo-hypoparathyroidism, he didn't seem to have other features that are inconsistent with Albright hereditary osteodystrophy. He had relatively intact cognition, was a bit short, so we took a look at him and thought that his pseudo-hypoparathyroidism, in part because he's one of more than eight siblings, the only one of which has symptoms, might be sporadic. And indeed on clinical testing for methylation at the GNES locus was found to have a defect in methylation consistent with sporadic pseudo-hypoparathyroidism type 1B. We further sent his samples to Dr. Dripner's lab at Harvard who did not find evidence for a deletion and did see complete loss of methylation in GNES as is typical in many patients with pseudo-hypoparathyroidism of the sporadic type. We also were curious about his history of progressive bronchiectasis and work closely with other colleagues at Stanford who have developed new assays to interrogate the metagenome. This highlights a few figures from one of Dr. Quake's papers showing the ability to detect both new and known microbes in circulating cell-free DNA. As you may have heard of circulating cell-free DNA as it's now reached common practice for non-invasive prenatal diagnosis, the challenge with pathogen detection is of the 30 billion DNA fragments per milliliter, only 1,000th of a percent are pathogen DNA. But with appropriate amplification protocols, there's ability to detect over 1,200 DNA microorganisms, DNA viruses, fungi, proso-aminatodes, and bacteria from blood. And so we used this technique and found that he, in fact, had Klebsiella pneumonia colonization. This was detected on two independent samples and was confirmed by a culture from bronchoscopy. There are no other significant microbial DNA detected responsible for his chronic lung disease. And importantly, his blood cultures drawn contemporaneously were negative. And the fifth case is a two-year-old male with history of failure to thrive, global developmental delay, hypotonia, and dysmorphic features including posteriorly rotated ears, prominent nasal tip, as well as two, three toe syndactyly, abnormality of the palmar creases, and low height and weight or age. He had undergone a prior genetic testing including normal prenatal testing in terms of trisomies, a normal chromosome microarray, normal Fragile X testing, and non-diagnostic proband-only whole-exome sequence. So in addition to whole-exome sequencing available through the UGM as we've seen in the second case, there's also availability of whole-genome sequencing. So he underwent whole-genome sequencing along with his parents. And this was notable for a novel copy number variant identified in ARID1B. This was absent in 592 whole-genome sequencing samples throughout the Undiagnosed Diseases Network. You can see the IGV plots here showing both the split reads and the reduced depth in this area of a 5KB deletion in the ARID1B gene that's not present in either parental blood sample. This was confirmed on independent clinical tests showing by microarray heterozygous deletion in ARID1B was thought to be likely pathogenic for Koffin-Cyrus syndrome, which he had overlap with several but not all of the features between himself and the typical presentation. So in summary, we've shown five patient stories and shown several different ways where a breakthrough sort of outside the Cummings standard of care exome sequencing has informed a diagnosis. And the first through deep sequencing looking for mosaicism, identifying a recently discovered disorder in terms of the molecular mechanism through data analysis, case matching, untargeted metabolomics, and a model organism study identification of a novel disorder through methylation testing and metagenomic assessment to have likely more than two diagnoses in the same patient through exome sequencing informatics reanalysis and confirmatory RNA-seq, demonstration of misclassification of a previously identified variant, and through whole genome sequencing and structural variant analysis, identify a novel small deletion. So in summary, the cases have, I think, reemphasized that certainly our field has been transformed by the use of whole exome sequencing, but there continues to be a frontier beyond which we can take things further, and this goes in different directions. Some of the categories of cases that today's talk, as well as our collective experience have emphasized is that we're not consistently and routinely identifying a cryptic splice variance. There's lots of opportunity in the area of non-coding variants, very small CNVs beneath the resolution of microarrays continue to be an area of opportunity, and mosaicism is an area of great potential. Specifically, I think that we readily recognize the cases where there are overt on imaging or on cutaneous findings, evidence of a flashcoid distribution. I think that it's likely that there are a significant number of cases where it's, so to speak, not visible to the naked eye and we'll collectively be developing strategies to think about how we might be routinely sampling multiple tissues in cases for calcitin to exome analysis. Case matching is essentially the heart and soul of the diagnostic enterprise, and as we in the network and others are currently working to develop systematic ways of documenting information and sharing it routinely with others across networks and across clinic sites, I think that we'll have all the more opportunity to increase our ability to reach diagnoses on a more consistent and expeditious basis. When dealing with very rare diseases, especially the initial identification of new disorders were by definition dealing with small numbers of patients, and in these cases, in addition to case matching, sorry, case matching being very important, functional studies are critical. In the cases we looked at today, RNA sequencing, metabolomics, and evaluations in model organisms were very useful. So that concludes our case presentations. We know that the presentation today will be made available as a durable educational resource to our community, so we have included references highlighting the UDN, some of the core bioinformatics tools that hold our network together, the phenotypes work in particular, references about the calling of copy number variants from whole genome sequence data on the metagenomics, and also the publication that reported the first association between a GNAQ and a phagomatosis pigmentovascularis. We also would like to take time to acknowledge again our network, our support through the National Institutes of Health, as well as locally here at Stanford, the Center for Undiagnosed Diseases, and the departments including but not limited to medicine, neurology, and pediatrics. We'd like to open it up for questions now. So, there's a question from Greg Lennon. Does none of these patients, mutations appear to have been submitted to ClinVar or updated in ClinVar, and will these be submitted to ClinVar? So, the UDN has a coordinating center that houses data and also bundles submissions to both dbGaP of sequencing data and to ClinVar of clinically interesting variants. Variants that were discovered through the sequencing cores are also deposited through standard mechanisms at those cores. So, those are things that, as we have increasing levels of confidence in particular variants, they are sent to the coordinating center and then submitted to ClinVar in batch format. The second question, for case three, there's RNA analysis was performed, and to see the effect on RNA splicing, was REB approval required before doing RNA analysis? So, by REB- Yeah, so there is an overarching central IRB for the UDN and the patients who, and their physicians who apply to the UDN. The patient agrees to a consent form for consideration of application and then a more in-depth consent form for evaluation and analysis of research data as is implied in the question, RNA sequencing and splicing analysis is not within the auspices of what is done in Clio labs to date. So, these were done on a research basis and the IRB and protocol and consent form have fairly broad support for pursuing research tests to increase the likelihood of developing a diagnosis for patients. Certainly, in certain cases, such as the transcriptomic case, then the challenge of the ability to reproduce that finding in a Clio lab test is something that across the network we try to achieve, but it continues to be a gray area between clinical work and research activities. Ultimately, what's returned to the patient is founded on the Clio laboratory-based testing and as in any case where there is a novel variant that's not truncating, there has to be an element of caution in counseling patients about such results because ultimately, until it's seen in more than one individual, one cannot be completely certain, but this is something that is true of really all missense or splice variants that are not necessarily at the plus one or plus two position. So, it's something I think that we all encounter in clinical practice and we navigate the uncertainty intrinsic with that and these boundaries between what's clinical on a full, thorough basis and what is at the margins in the research realm. And then a third question, and we welcome more, is related to the patient with the GNAS methylation defects. I'm gonna switch to the slide showing both the clinical tests and the research tests done to look at that. So, you see on the right side of the slide that we have the different isoproteins from GNAS and then the clinical test at the Johns Hopkins lab is looking for exon 1A loss of methylation and then the research test is looking at loss of methylation across AB as well as the AS and partial loss of methylation at the XL locus denoted by the light blue arrows. So, certainly there's a review and nature reviews and immunology that goes into great detail about that particular locus and differences in methylation at that locus. And then we have another question asking about the tissue used to identify the mosaic case. So, there were actually a number of skin biopsies were taken from the patient's back and some were taken from what was visibly typical appearing skin and some were taken from the area that appeared to involve the nevus flamius. And so, they were skin samples and they were from what appeared to be affected versus unaffected areas of the skin. And even in the areas that appeared to the eye to be affected, it was a minority of the sequencing reads, but still it was 5 or 10% I believe that they were evident there. So, it required a skin biopsy and it requires sequencing to considerable depth to look in the DNA from the uncultured cells. So, DNA was extracted directly from the biopsy in order to initially identify that pathogenic variant. And as Dr. Bernstein mentioned, contemporaneous with us doing that work is a definitive paper showing association with this particular variant that was found with many of these cases was published. All right, if there are, we're happy to take a moment for any other questions, we appreciate them. If not, I think we have one or two more closing messages from the college. So, here is the web link for where today's question and answer session and the presentation will be available at acmg.net. And then finally, another reminder that everyone is looking forward to the meeting next spring and abstracts can be submitted for the next several weeks online. Are there any other questions? If there are, we'll be here for another 19 minutes and happy to answer questions. So we have another question, will all institutions use the same platform for testing? This is from Dr. Amarillo at Wash U in St. Louis. So there is, we'll go back to the network diagram. So there are two sequencing cores, one at Hudson Alpha and one at Baylor College of Medicine. So in terms of sequencing, there's whole genome sequencing currently done at Hudson Alpha and then whole exome sequencing done through Baylor Genetics Laboratory in Texas. And so they both are using Illumina short reads and Baylor Genetics Laboratory is using their custom capture. They each have different calling and interpretation pipelines and they both are able to provide clear validation of variants that are identified through sequencing. Thank you. We have a question asking for some more detail about the mitochondrial case, and if we can go back to that one. So there's more that can be said about the program's history as well as to the second patient that was identified. So actually, the two patients had different histories in some interesting respects. So the program presented at two days of age with a metabolic crisis, and it's had more than a dozen of such episodes in her first, I think it's now eight years of life. Over time, it's been clear that they've generally been responsive to IV fluids with dextrose with a possible response to a mitochondrial cocktail of cofactors and antioxidants. Very difficult to say that with N of 1. She has had some very mild developmental differences, but is at appropriate school grade for age, and she is of slightly short stature. As she's gotten older, she has had evidence of easy fatigability and not necessarily frank weakness, but does have difficulty walking with distance. The second patient that was identified in the UK presented later, so I believe at school age, presented during an acute illness with encephalopathy and was hospitalized in ICU. So it's had a far wider period of time between episodes that perhaps are not so easily provoked by illness. And this may relate to a different, in a sense, mutation that was present in that patient. It could also be, of course, other factors. And so far, it's only the two cases, and this was presented at the ASHG meeting preliminarily last year and will be submitted hopefully soon for publication. So with time, maybe we'll find additional cases and more fully understand the spectrum of this disorder. In the patient, the local patient, there has been evidence of recurrent muscle breakdown as evidenced by increased creatine kinase in blood. This has not been present in the other patient. Both patients have demonstrated the 3-methylglutakonic aciduria. I can also say a little bit more about what's been learned about phycomatosis stigmata vascularis from the patient we presented this afternoon, as well as from the very small emerging literature on this condition since the denomination of the molecular basis. So historically, the condition of phycomatosis stigmata vascularis was divided into syndromic or non-syndromic cases, and it's more recently realized that it's indeed a spectrum. So, AAPL proposed that it could be divided into three subgroups based on the type of the kinase manifestations, the vascular and the pigmentary differences, and the nature of the vascular-pigmentary differences. So that's Spilla-Rosia, Cesio-Marmorata, and the third one has Amoebis Spillus component. The first two, the Spilla-Rosia and the Cesio-Marmorata, have been connected to the GNIQ mutations as of recently the third class has not, so we'll see at a molecular basis if these are indeed part of the same spectrum. We've been very interested in the central nervous system manifestations of the condition, so a number of patients, especially those in the distribution similar to Sturge-Weber, in the first or second branch of the trigeminal distribution, have had seizures, intellectual instability. We are not aware of other patients that have had a movement disorder as seen in our patients, so we're interested to learn if this is maybe something that is present in other patients or could be something occurring independently of his Mosaic-GM-AQ disorder. We're working on a review of cases of patients with this family of conditions with a focus on the more recently molecularly defined cases, although because it is as highlighted in the questions a Mosaic disorder, many patients don't undergo tissue sampling other than blood, which enables that, so it may be slower progress. Are there questions in more detail about any of the other cases? So, for our adult cases, I think one of the important distinctions is often we have to look at the certain things that are acquired versus things that are congenital, but they're actually delayed onset, and so in addition to the metagenomic work done by ourselves and others, there's also extensive environmental history questionnaire that is offered to all of the participants in the Undergrad Institute's network, and this is coupled with an environmental exposures working group, which provides reviews of this questionnaire and then opportunities for occupational or environmental medicine evaluations and to link potential environmental exposures to findings, particularly in our adult patients. In addition, patients such as this one who have undergone colorectal sequencing but have not yet found a diagnosis, we submit promising candidate variants both to the model organism screening core for consideration and then through matchmaker exchange through the world at large to attempt to find matches of novel or very rare genetic variants. Obviously, a challenge in some cases, including in our Filipinos, is relative paucity of normative data, and so efforts across the Undiagnosed Diseases Network and Undiagnosed Diseases Network International to facilitate data sharing can really aid in discovery and diagnosis in patients regardless of their ethnicity. So, those are some of the challenges of our adult patients. I would also say that this patient, like many others, both within our center and across the network, remain undiagnosed despite hard work by many of the people within the network. So, there is certainly more to be learned. I think that metabolomics is an area where we really are at the front end of the curve in understanding typical variation and variation within disease. We certainly have a couple of cases where it's been very informative, but it's quite an exercise in just defining the range of normal and how that's impacted by diet and medications and other exposures. The other area there is there's far more infrastructure for the efficient and ready sharing of molecular genetic data than there is for metabolomic data, and with time, hopefully, there's opportunity for that to improve as well. So, related to metabolomics, there is a question, when do you use metabolomics and how do we choose patients that could benefit or where it might aid in the diagnosis? So, I think that's clearly something that's in the NICU. I think there are suspicions based on historical experience with more of what's now traditional biochemical profiling in terms of organic acids and isocarotenes and amino acids. People probably sacrifice and such. And so, there's been proportionally more success in neurologic disorders. For example, to show a progressive nature or an episodic nature. The expectation is that what is informed by metabolomic analysis will be broader than what we've seen through a sort of more limited scope of the metabolome. And we're all learning about this together. In parallel to our work in the UDN, we're going out here and other institutions is looking across a wider spectrum of disease than conventionally has been thought the purview of biochemical genetics to see if there may be metabolic signatures of other families of disease, including organ failure, perhaps various inflammatory states may have a characteristic metabolomic profile. So, hopefully, we can optimistic that there's a lot to learn in this area. But for right now, we're collecting the samples and banking them for all of our patients that can be used for metabolomics. And we're putting a priority on undertaking a metabolomic analysis for patients with neurologic, especially intermittent or progressive neurologic findings. And I think in addition to thinking about positive metabolomics, there's a growing interest in fecal metabolomics and also CSF metabolomics as well. And there is a comment about the network. And so, we are anticipating the second phase of the UDN to start sometime next year, which will help expand the national network. There's also an international network, the UDNI, which encompasses not just outside of U.S. undiagnosed diseases centers and programs, but also centers and programs that have similar missions to the Undiagnosed Diseases Network that NIH sponsored within the U.S. So, anyone who is interested or has developed a program similar to aspects of the UDN is welcome to look into the UDNI as an opportunity for the collaboration and sharing in these diagnostic challenges. We have a perceptive question coming in at Children's Mercy Hospital. And the question is, how central to establishing a diagnosis for the metabolomics in case two? And I think it was a collateral piece of evidence. I think that the principle evidence was the genetics and the two independent families, so one in the U.S. and the U.K. I think were, from the geneticists, the strongest evidence. And then, the functional studies in Drosophila were, I think, additionally quite compelling. And then, the biochemical picture, I think, is important. It was a significant part of the phenotype. Some of that is just, for example, the lactic acidosis, the 3-methylglutakonic aciduria, important. And I'm sort of fleshing it out in additional detail. I think it was helpful to have the metabolomics because it highlighted certain pathways in more detail that have been evident from the clinical studies. Some attempts are still in progress. We may actually compare the metabolomics from the fly, the Drosophila model, to the human patients. And it raised a lot of interesting questions about the limitations of sampling plasma versus when you're analyzing metabolomics of a fly. It's essentially a whole-body assay and how metabolomics can be adapted or performed between model organisms and humans and what the best approaches are for that. So the concise answer is, no, it was certainly not the whole story, but I think it's an interesting part of the story and one that gives us ideas and information that will be helpful in future cases. And I think, equally, an attempt to look for other cases across the network with a similar metabolic profile did not find other cases through the metabolomics core that were similar in nature to this case. So certainly one could imagine in a different case you would do that and discover two cases that are actually much more similar than you might otherwise think. I think it's going to take time in looking at a far larger number of patients. If we're thinking about rare and especially distinctive profiles, it's probably going to start to take thousands of patients or tens of thousands or more. And these are very early days, I think, in these metabolomics investigations. Great. So again, we'll just highlight if you have any need to review what was said and shown today. The recording will be available here. And then do submit your abstracts by December 8th so that you can participate in the ASMD clinical meeting. And we want to thank everybody for participating and appreciate the opportunity to show some of the opportunities through the Undiagnosed Diseases Network.

Undiagnosed Diseases Network

clinical site

diagnostic science

rare diseases

collaboration

data sharing

genetic variants

diagnostic process

improving diagnosis