Backed by Computer Power, Scientists Are Finding the Causes of Mysterious DiseasesAug 8, 2016
Some diseases are so rare and unusual that doctors have never seen anything like it. An excruciating journey for both families and doctors, figuring out what’s wrong can take years, if an answer is ever found at all. Using a computer tool developed by Aaron Quinlan, Ph.D., he and his team recently uncovered the genetic causes behind nearly one dozen previously unsolved cases, all infants with severe seizures. Quinlan describes GEMINI, and how it is helping he and his colleagues at the USTAR Center for Genetic Discovery to understand rare disease genetics.
Interviewer: With the power of computers behind them, scientists are solving the mysteries of undiagnosed diseases, up next on The Scope.
Announcer: Examining the latest research and telling you about the latest breakthroughs, the Science and Research Show is on The Scope.
Interviewer: I'm talking with Dr. Aaron Quinlan, associate director of the USTAR Center for Genetic Discovery at the University of Utah. Dr. Quinlan, you recently had some really exciting results using technologies that your group developed. They may have helped solve a health mystery. This is about infants with a particular condition. What was going on?
Dr. Quinlan: We were studying infants with seizure disorders, and the genetic basis of those seizure disorders was unsolved.
Interviewer: So, the idea is that . . . I mean, obviously they had seizures, presumably pretty severe ones, but doctors didn't know what was causing it. So, there were about a dozen or cases, and you were able to possibly find the cause for most of them?
Dr. Quinlan: Yeah, for the majority, I guess 90% of the cases we have a pretty clear candidate that we feel strongly about, and in 9 or 10 of those cases, it's a mutation in a gene that is known to cause this phenotype but was not picked up via standard clinical diagnostic tests, and in a handful of other cases, we think we have discovered new genes that underlie this phenotype.
From a clinical perspective, there's a transition, certainly removing very rapidly from gene panel tests, where we only look at a very, very small subset of the genome to interrogate genes that we know cause a given disease phenotype to, I think, in the coming years, it will be a standard course of care to use exome or genome sequencing to do this diagnosis because it's so effective, and I think the clinicians that we were working with were very excited about the accuracy and the rapidity with which we could make these predictions.
Interviewer: The role of you and your group in this is that you've developed a computational tool called Gemini, and that's what led to these results. What is Gemini?
Dr. Quinlan: So, we used genome sequencing of both the infant and their parents to try and identify genetic mutations, essentially, that cause the disease phenotype in question, and this process requires a broad spectrum of computational methods, everything from rapidly and accurately processing the sequencing data to identifying genetic variants that exist in these families, and then finally to essentially get back to a needle in the haystack problem of what is the single genetic mutation that causes the phenotype and isolate that from the potentially millions of genetic variants that are benign but exist in these infant genomes.
So, the idea is that Gemini takes all the genetic variation that's observed in the genomes or exomes of all the individuals that you're studying, and it integrates all that genetic variation information with the extreme wealth of genome annotations and reference databases that we have. For instance, some people might be familiar with OMIM. It's a list of all the known mutations or genetic variants and genes that are associated with diseases.
Interviewer: Right, so keeping up with the pace of research, the pace of knowledge.
Dr. Quinlan: Right. It's an incredibly demanding problem because there's probably 50 to 60 reference databases that we try to use, and they're all evolving. They all have mistakes. Those mistakes are fixed, and you've gotta propagate those fixes to the mistakes as quickly as possible so that . . . what we're trying to do here is empower discovery for human genetics, and so, having the latest and greatest information, obviously, empowers that process.
Interviewer: So, is there somebody who's monitoring each of those databases and saying, "Oh, gotta update, gotta update, gotta update"?
Dr. Quinlan: Yeah, we have people in the lab who monitor that, but, believe me, the research community that uses this software, they monitor it as well.
Interviewer: And so, the real tricky part is that a lot of us have scads, you can give me the numbers, you know, scads of variations in our genome, and so that the problem is finding the one or ones that increase risk for a certain disease.
Dr. Quinlan: That's right. I mean, any two individuals differ by about 3,000,000 to 4,000,000 genetic variants. So, when you look at a family, do a whole genome sequencing of an entire family, you're going to find on the order of 3,000,000 to 10,000,000 genetic variants that you have to sift through. Now, many of those, admittedly, are very simple to ignore, especially for rare disease phenotypes. We typically focus on genetic variants that affect protein coating genes. But even when you do that, you're talking about on the order of 18,000 to 20,000 genetic variants that need to be considered, and so, we need to be able to do that in a quick and reproducible way, and we want to minimize false predictions, but I think even more concerning are real genetic variants that may be associated with the phenotype that you miss. So, we want to essentially find everything but don't over-predict.
Interviewer: I imagine you spend a good part of your day in front of a computer screen. I'm wondering do you think about how this sequence of letters you have in front of you is actually a real person.
Dr. Quinlan: Yeah. Admittedly, I am fairly disconnected. I'm a genetic researcher that spends 12 to 15 hours a day in front of a computer, and I'm not a clinician, so, I don't interact with patients on a day-to-day basis. However, I mean, that is our motivation here, is, you know, that was the main reason I moved my lab from the University of Virginia to the University of Utah was to have that connection.
We have a very nice interaction between researchers and clinicians here at the U, and I think it really helps to bring home the reality of these cases. We meet with the doctors who actually work with these patients, and when you understand their plight both in terms of the diagnostic odyssey and also the impact on these families, both in the short and long term, it makes it very real.
I would like to be able to provide a resource to try and solve rare disorders in Utah, nationally, and not only retrospectively for families that are sort of pursuing this diagnostic odyssey, but also to have a system where this can be done in real time in collaboration with clinicians in our hospital and other hospitals so that when there's an infant that comes through the NICU or there's some pediatric genetic disorder that is perplexing, we have a system in place where we can sequence the genomes and actually bring our tools to bear on solving that problem quickly and as accurately as possible.
Announcer: Interesting, informative, and all in the name of better health. This is The Scope Health Sciences Radio.