November 18, 2015
As the mysterious links between genetics and disease continue to reveal more of their secrets, Hertz Fellow Hilary Finucane is right at the forefront of improving scientists’ ability to extract biological insight from large genetic datasets.
Finucane, a PhD student at MIT, has been busy lately, as her involvement in the publishing of two recent papers in the journal Nature Genetics can attest. In broad terms, the papers allow researchers to learn about which cell types and tissues are important for particular diseases, and also which pairs of traits and diseases have a common genetic basis.
The methods they devised, called stratified LD score regression and cross-trait LD score regression, only require the use of summary statistics. In genetics, many of the largest datasets are split across many research groups, each of which has collected data on several thousand individuals, but researchers cannot share this individual level data, often for privacy reasons. Instead, they make available a condensed summary of the data they have collected. By requiring only summary statistics, which can be shared, the LD score regression methods can be applied to large sets of data that were not previously amenable to this type of analysis.
“There are other methods that work quite well, but they require institutions to share individual-level data. We created a new set of methods that enable analysis on a larger scale by requiring only the summary statistics,” Finucane said. “The methods allow scientists to get the benefits of large sample size even when no single researcher has access to all of the individual-level data."
Finucane was a lead author on the papers describing these two LD score regression methods, along with co-authors and colleagues Brendan Bulik-Sullivan, Benjamin Neale and Alkes Price. The method is catching on with research groups faster than she anticipated.
“These are people who are working on the hardest problems in genetics,” Finucane said. “It’s been a nice surprise how strong the connection is between the people who use the methods and the people who create the methods. It’s becoming a standard.”
Finucane, 28, grew up in Columbia, Maryland and is the daughter of two physician parents who encouraged her pursuit of science. She majored in mathematics at Harvard University, but switched her focus to computational biology at MIT because she wanted to more directly see how her research could benefit people.
Since starting her PhD in 2012, much of her work has been in attempting to understand the relationship between pairs of traits and diseases, and discovering hints on the causes of common heritable disorders. While scientists still don’t know how much of the triggers for disease are environmental versus genetic, she said, genetics can still help scientists plan the next stage in understanding the causes of disease.
“In some traits, we’ve made a lot of progress, but the traits I like to study are the least understood,” she said. “I’m hoping this research will be one small step in a very long journey towards curing these diseases.”
Finucane, who has about a year left on her PhD, said she’s interested in remaining in academia and perhaps becoming a professor. She lives in the Boston area with her husband Yakir Reshef, an MD/PhD student at Harvard and an occasional collaborator.
Graphic courtesy of Chrysos Designs: Some of the Connections Identified Using the New Methods.