Happy hunting with Orion
By Stacy Lawrence
Elucidating the role of genetics in understanding disease is a driving force for the diagnostic and biopharmaceutical industries. But, until now, much of the human genome has remained largely a mystery. Scientists have long focused on the portion that codes for proteins but that represents only 2 percent of the human genome.
Researchers at Columbia University Medical Center have developed a new technique, known as Orion, which interrogates the noncoding 98 percent, the so-called 'dark genome.' They have released it to the public and expect it to be immediately useful in discovering the genetics related to the roughly two-thirds to three-quarters of suspected genetic disease for which the cause is unknown.
"What we did here is really to develop a framework. And I think we showed that the framework works as intended. Basically, we had already known from our earlier work that you can reasonably score genes in ways that relate to whether the genes cause disease or not; that's already been proven very, very helpful for interpreting variation in individual genomes," lead author David Goldstein, the John Borne professor of medical and surgical research and director of the Institute for Genomic Medicine at New York's Columbia University Medical Center, told BioWorld MedTech.
"What we wanted to do is just ask the question Can the same general approach be applied outside of genes? That's what we've done with this Orion methodology and what the paper shows can be done. We have a variety of evaluations in the paper that show that the basic approach works and is applicable outside of genes," he concluded.
Genes, and the exomes based upon them, are inherently part of the process of genetic coding that's already relatively well understood.
Orion was developed by identifying regions in the noncoding genome that were constant across the 1,662 sequenced whole genomes that the researchers examined. The idea is that since these are highly consistent, or intolerant of mutations, these regions must be key to healthy human function and any variations in them could be pathogenic. Similar studies have been conducted previously across species, identifying intolerant regions shared by several kinds of animals.
"We show that Orion is highly correlated with known intolerant regions as well as regions that harbor putatively pathogenic variation. This approach provides a mechanism to identify pathogenic variation in the human noncoding genome and will have immediate utility in the diagnostic interpretation of patient genomes and in large case control studies using whole-genome sequences," concluded the paper that was recently published in the open access journal PLOS One. Goldstein's Columbia colleagues Ayal Gussow and Andrew Allen also worked to develop Orion.
Big data boost
Although, they expect Orion to be immediately useful to other researchers seeking genetic disease linkages that have eluded prior techniques, the researchers also have big plans to improve Orion that are expected to be executed in the next year. The Columbia researchers are already working to incorporate massive volumes of sequenced whole genomes into Orion to better define the regions intolerant of mutations, as well as the pathogenic variations that can occur.
"So, what we need to do is to take this basic approach, which we now know works, and apply it to 50,000-100,000 whole genomes and get a much more accurate description of which parts of the human genome are indeed under selection," said Goldstein.
"At that stage, I do believe, it will help us to identify mutations outside of genes that influence risk. That's the next step for us, and there are data sets of that size around now. This really just goes back to the point that genomics works best when you practice it on a very large scale," he added.
The paper offers a series of intolerance scores for given noncoding genomic regions. Goldstein expects incorporating such a massive whole genome dataset will enable a new, more accurate set of these scores. The researchers plan to test the new scores by comparing them with whole genome sequence data for patients who are expected to have causal mutations that don't appear in their exome.
Dx, biopharma implications
Goldstein expects that Orion will be useful in identifying noncoding genetic mutations that are suspected to play a role in disorders such as autism and epilepsy. The paper already found known de novo mutations present in autism and epilepsy patients that fall into the intolerant regions they have identified.
The application of Orion is expected to be useful in identifying the underlying mutations related to genetic disease; in most cases where there is a suspected genetic link, it can't be made to the coding genome. But it's not clear yet what this technology will mean for drug development as biopharma researchers are ever-more focused on understanding and tapping into genomic dynamics.
"The better we get at doing genomics, the better we can use genomics to inform drug development," said Goldstein. "This will help to analyze any genomic data where you're trying to relate variations to phenotype, outcomes that predict disease predisposition or drug response. But I don't see it being immediately transformational in that context. Whole genome is a fair bit more expensive than whole exome sequencing."
Goldstein envisions that this research could lead to major advances, such as the ability to screen newborns and even adults to assess the risk of expressing a genetic disease and to prepare for, or even work to prevent, that outcome.
"I would like us, within roughly the next 10 years, to really get good at recognizing the genetic differences amongst people that do something. Right now, we are reasonably good at looking for a mutation that explains a severe genetic disease. We are reasonably good, but we could get better at that," he said, adding, "A third of the time you find a mutation that is responsible, and that's pretty powerful. So, we're pretty good at that, and we'll get better."
Published August 11, 2017