Blog @ Illumina
Real scientists. Real commentary.

A Sequencing Jamboree: Exomes and Genomes

Julie Collens
| Nov 13, 2012

Over the years, I’ve looked forward to ASHG as a chance meet up with old colleagues and make new friends, events largely facilitated by post-session parties and get-togethers. This time, I attended a different kind of celebration: the session titled “A Sequencing Jamboree: Exomes and Genomes.” Talks in this session discussed novel approaches to identifying causal variants and gene regions responsible for a variety of diseases, including type 2 diabetes (T2D), high LDL, cystic fibrosis, and autism.

diabetes ribbon

The first talk by Teslovich shared data from the T2D-GENES Project of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), which plans to perform deep exome sequencing with 10,000 samples from five ancestry groups. The group shared variant data for the 5,300 samples sequenced so far. They classified variants into three categories: not likely to be deleterious (synonymous), moderately deleterious, or predicted to be most deleterious (nonsense, frameshift, or those occurring at essential splice sites). By comparing variants by ancestry group, they found that 5% of all variants are shared by all groups, and 84% by a single ancestry. They also found that the variants predicted to be most deleterious were the smallest proportion shared by all ancestry groups, consistent with the hypothesis that most deleterious variants are rare and observed in a single population. From there, they used single marker analysis to look for T2D-associated variants specific to ancestry groups. The top hit was found in the East Asian samples in PAX4 R192H, a gene essential for the development of pancreatic islet cells.

I also enjoyed Susan Walker’s talk on identifying susceptibility genes in Autism Spectrum Disorder, which is characterized by a collection of disorders and rare variants. Her group sequenced the exomes of 90 unrelated cases with de novo or rare inherited CNVs and looked for additional variants. Walker mentioned that their group would like to move to whole-genome sequencing—as the exome excludes many non-coding variants that they suspect may also be involved in the disorder—and advocated for an approach that combined multiple technologies to describe the genetic architecture of genomic variants. Though the study characterized unrelated individuals, she concluded that the impact of variants is best interpreted at the individual or family levels, but population studies are required to generalize interpretations.

Compared to previous years, this ASHG meeting revealed much more interest in looking at traits in a population-specific and admixed context, as well as a more sophisticated approach to analyzing rare variants in populations and samples. Approaches converged on the hunt for disease-causing variants using families, single ancestry, or diverse ancestry options. I also frequently heard the call to develop population-specific reference genomes or panels to provide higher power for finding population-specific variants, even when using much smaller sample sizes. These focuses suggest that exome and genome sequencing will continue to be instrumental in shaping new methods for identifying causal variants for diseases.