We’ve been big fans of the structural variation work coming out of Evan Eichler’s University of Washington lab for years, so when he was the subject of a Mendelspod interview, we couldn’t wait to listen. If you missed the podcast, check it out here.
In his chat with Mendelspod host Theral Timpson, Eichler begins with an overview of structural variation and a really interesting perspective on how duplications were handled during the Human Genome Project by both the public and private initiatives. It was a nice reminder that structural variants have dogged the genomics community from the beginning!
Eichler’s career has included a focus on segmental duplications, the hotspots of rearrangements that tend to be home for most of newly acquired genes that make us uniquely human. These areas are replete with structural variants, which is why Eichler put such effort into resolving those elements. The segmental duplication regions are dynamic, frequently implicated in disease, and “move at light speed compared to most regions of the genome,” he tells Timpson, calling them “crucibles for evolutionary change.”
SNPs, Eichler says, are kind of like tremors in the genome, while structural variants are full-blown earthquakes. He notes that while NGS technologies allow us to observe SNPs directly, we’ve had to infer structural variants from this data, which limits our ability to detect these variants accurately and comprehensively. While he believes that short-read sequencing was revolutionary for genomics, he made the case that for optimal clinical utility, we have to move toward a de novo assembly approach for each person’s genome that will fully phase maternal and paternal haplotypes.
Eichler is looking to long-read sequencing to discover and catalog existing structural variation in humans as well as to produce de novo assemblies. This technology can uncover 90% of the structural variants that have been missed by short-read sequencers, he says. Today, long-read sequencing is too expensive to replace short-read tools, but Eichler suggests that could change. “I think if we had a long-read sequencing technology that was even double the price of Illumina, short reads would be dead,” he tells Timpson. But in addition to lowering costs, he says, these technologies have to produce megabase-scale reads while also increasing capacity in order to make these platforms better suited for clinical use.
Eichler also sees room to improve sensitivity and characterization of structural variants. “What we need to think about is how to do this right, and that means understanding all the variation from stem to stern in these genomes,” he says.