Research Roundup: May 29, 2020

Human variation surveyed, polygenic prediction enhanced, Levantine history explored, and more

Kelly Davidson
Credit: Kelly Davidson

Welcome to the May 29, 2020 installment of Research Roundup, a recurring snapshot of recent studies published by scientists at the Broad Institute and their collaborators.

Surveying the landscape of human genetic variation

For eight years, the gnomAD Consortium — including scientists from Broad's Program in Medical and Population Genetics (MPG), Data Sciences Platform, and Genomics Platform, as well as more than 100 investigators internationally — has been compiling what is now the world's largest publicly available catalog of human genetic data: more than 125,000 whole exomes and 15,000 whole genomes. This week, in a group of seven Nature, Nature Communications, and Nature Medicine papers, consortium members described the first set of discoveries to come out of this massive dataset. Together the papers present the most comprehensive survey to date of human loss-of-function (LoF) and structural variation, show how tools that account for variants' context can aid their clinical interpretation, and illustrate how population-scale LoF variant data can help evaluate proposed drug targets. Learn more in a Broad news story, a Nature editorial, and coverage in Cosmos, El Mundo, and AFP.

New role for a tumor suppressor in immunotherapy resistance

The response to cancer immunotherapies, such as PD-1 inhibitors, can vary dramatically, even among a single patient’s tumors. A team led by Cécile Gstalder and associate member Rizwan Haq of Dana-Farber Cancer Institute characterized tumors from one melanoma patient, finding that a resistant tumor carried a loss-of-function mutation in the tumor suppressor gene FBXW7, while a drug-sensitive tumor did not. Inactivation of the gene in murine models caused resistance to anti-PD-1 therapy, and loss of the gene impaired the signaling pathways that respond to viruses. Described in Cancer Discovery, the study suggests that therapeutic reactivation of these pathways could help some patients.

Insights into liver metabolism 

NADH and NAD+ are essential redox cofactors that play an important role in hepatic metabolism. However, there is still a limited understanding of the role of NADH/NAD+ ratio in metabolic physiology. Broad Metabolism Program’s Russell Goodman, institute member Vamsi Mootha, and colleagues used a bacterial genetic tool called LbNOX combined with metabolomics to characterize the biochemical effects of lowering the NADH/NAD+ ratio in mice. The authors identified circulating α-hydroxybutyrate levels as a robust biomarker of an elevated hepatic cellular NADH/NAD+ ratio, which is in turn influenced by common human genetic variation in GCKR, a gene associated with fatty liver disease and diabetes. Read more in Nature.

New polygenic risk prediction method for common diseases 

In the American Journal of Human Genetics, Sung Chun, Maxim Imakaev, MPG associate member Shamil Sunyaev, Nathan Stitziel (Washington University), and colleagues propose a new polygenic risk prediction method called non-parametric shrinkage (NPS) that does not require explicitly modeling any underlying genetic architecture for each disease to be predicted. The researchers evaluated the performance of this new approach under a simulated genetic architecture of five million dense single nucleotide polymorphisms across the genome. They also tested the method using real data in four disease areas: breast cancer, type 2 diabetes, inflammatory bowel disease, and coronary heart disease. NPS improved the identification of groups at high risk for these four diseases, all of which have available early intervention or prevention treatments.

G-PROBing multiple disease diagnoses

Rachel Knevel, institute member Soumya Raychaudhuri, and their colleagues have developed a tool that uses genetic risk scores to calculate the probability of different diseases in situations where multiple diseases are possible. The team tested the Genetic Probability tool (G-PROB) for inflammatory arthritic diseases, including rheumatoid arthritis, gout, and others in more than 1,700 patients in three biobank cohorts. They found that for all patients, genetics could be used to rule out at least one disease. In 45 percent of patients, a likely diagnosis was identified with a 64 percent positive predictive value. This is practically useful, since in 35 percent of cases, the clinician’s initial diagnosis was incorrect. Learn more in Science Translational Medicine.

Genomic history in the Eastern Mediterranean

A team led by MPG senior associate member David Reich, Liran Carmel (Hebrew University in Jerusalem), and colleagues assembled a genome-wide ancient DNA dataset of 93 individuals to explore the genomic history of the Southern Levant, a region now recognized as Israel, Jordan, Lebanon, the Palestinian Authority, and southwest Syria, dating from the Bronze Age to the Iron Age. They determined that these individuals descended from two primary sources, with a non-local contribution increasing over time. They also found that genomes of Levant-related modern populations are linked to the Bronze Age Levant population, along with other unclear ancestry sources. Read more about the team's findings in Cell, news releases from Cell Press and the University of Vienna, and coverage in National Geographic.

To learn more about research conducted at the Broad, visit, and keep an eye on