Clinical Genomics and Variant Interpretation Were Key Themes at AGBT 2015

AGBT 2015 was a great event, and the QIAGEN Bioinformatics team was honored to participate. We’re grateful to all of the scientists who took time out of a hectic schedule to visit our lanai and the software demo session!

Linking genetic variation to phenotype was a major theme of this year’s meeting — and as developers of tools to help scientists make sense of DNA variants, that was pretty exciting. In the opening session, David Goldstein from Columbia University spoke about a large-scale study of people with epilepsy that involved sequencing more than 350 trios of parents and affected children. They found a lot more de novo mutations than expected, Goldstein noted, adding that his team is working to group patients by variants to establish subtypes of epilepsy for better treatment outcomes. In the same session, Yale’s Rick Lifton told attendees there is plenty of room for new discoveries about genetic mutations, and that routine sequencing of patients will be important to improving healthcare. Steve McCarroll from Harvard used allele frequency cutoffs to filter de novo mutations found in exome sequence data from a schizophrenia cohort, which had us cheering because that’s the concept behind our newly launched Allele Frequency Community.

Later in the conference, Evan Eichler from the University of Washington spoke about the importance of generating a comprehensive catalogue of genetic variation — not just SNPs, but structural variants and other complex elements — in the human genome. Eichler’s team has worked diligently and methodically to sequence two haploid human genomes from hydatidiform moles to get closer to what’s being called a platinum-grade genome that would be an unprecedented resource for the community. Along the way, they have shown that structural variants are dramatically underrepresented in the current human reference genome.

Elaine Mardis from Washington University’s Genome Institute gave a great talk on a study that revealed mutations predictive of worse outcomes for patients with ER+ breast cancer, as well as an AML study that looked at genomes along with transcriptomes. In another clinical talk, Gail Jarvik at the University of Washington spoke about the challenges of interpreting incidental findings from sequence data. A cross-lab study she presented showed that labs interpret variants differently, with some labs reporting a variant of unknown significance while other labs call the same variant “likely pathogenic.” Jarvik called for sharing of better annotations so labs can base interpretations on richer data sets. Similarly, Stanford’s Euan Ashley reminded attendees that the community still needs standards, such as benchmarking information for variant calls, to make genome interpretation clinically useful.

Congratulations to the scientists who gave podium or poster presentations at AGBT! We look forward to another great event next February. In the meantime, we’ll keep busy making our variant interpretation tools even more useful for this community.