Getting tumor genes out of normal data, marking up the non-coding genome, epileptic variants, and more.
Research Roundup: June 29, 2018
Welcome to the June 29, 2018 installment of Research Roundup, a recurring snapshot of recent studies published by scientists at the Broad Institute and their collaborators.
Clearing out cancerous confounders
To find cancer-related mutations, scientists compare sequencing data from patients' tumors with those from nearby healthy cells. But tumor cells often spread into adjacent tissues, potentially obscuring important mutations by making them appear to be part of normal DNA. To help clear things up, a team led at the Broad by Amaro Taylor-Weiner, Chip Stewart, and institute member and Cancer Genome Computational Analysis group director Gad Getz developed deTiN, a computational tool that estimates the amount of tumor-in-normal contamination within "normal" sequence data and highlights DNA variations that might otherwise be overlooked. They detailed deTiN in Nature Methods.
A clearer view of disease’s genetic architecture
Many hits from genome-wide association studies are in non-coding regions of the genome and may influence disease risk by altering the activity or epigenetic marks of other genes. A team led by associate member Alkes Price, Farhad Hormozdiari, and Steven Gazal in the Broad’s Program in Medical and Population Genetics (MPG) developed an approach to annotate these so-called quantitative trait loci (expression QTL, histone QTL, methylation QTL, and splicing QTL) from rich databases such as from the Genotype-Tissue Expression (GTEx) project and the European BLUEPRINT consortium. Reporting in Nature Genetics, the scientists show that their approach localizes a large proportion of disease heritability into a very small proportion of the genome. These findings shed light on the genetic architecture of diseases and complex traits.
Getting specific with disease locations
Identifying the underlying causal mechanisms and genes of common variant associations with complex diseases is an ongoing challenge, as the majority of associations lie in non-coding regions of the genome. A team led by associate member Ayellet Segrè in the MPG, Eric Gamazon at Vanderbilt University, and Kristin Ardlie, director of the GTEx project at the Broad, presented a study in Nature Genetics examining how expression QTL (eQTLs) from a range of human tissues help characterize the landscape of potential causal genes for thousands of known trait associations and discover new trait associations. eQTLs are specific chromosome locations that explain genetic variance of gene expression levels, which may also be associated with complex diseases and traits.
Exploring epilepsy’s genetic variants
Epilepsy is a frequent feature of neurodevelopmental disorders such as autism spectrum disorder, but it’s not clear what genetic differences separate the cases that occur with epilepsy from those that do not. A team led by MPG postdoctoral researcher Henrike Heyne and the University of Leipzig's Johannes Lemke sought to answer this question by profiling 6,753 patients with different neurodevelopmental disorders, including nearly 2,000 individuals also diagnosed with epilepsy. The researchers identified new disease-associated genes with implications for routine diagnostics and therapy. Read the full story in Nature Genetics.