Joseph Pickrell: Methods to Infer History
Postdoctoral researcher, David Reich lab, Harvard Medical School
Recommended by Molly Przeworski, University of Chicago
Population genetics falls at the nexus of what Joseph Pickrell is interested in. It contains elements of statistics and computation as well as human history, such as being able to address questions about how people have been able to adapt to living in very hot and very cold environments.
"I just was never very good at biology and I really loved statistics and computation and things like that. And so it was a natural place for me to end up," Pickrell said.
Currently, Pickrell is working on developing computational tools to aid in analyzing population history data, particularly for regions like Africa where there are complex population histories. In Africa, people have moved around a lot, mixed with other populations, and adapted to the very different climates found across the continent.
At the same time, Pickrell is also working on squeezing the most out of genome-wide association studies to link their findings to biology. For example, from GWAS on height, he said that he would like to tease out "the biology of why I'm the height I am and someone else is the height they are." Similar work, he added, could be done on diseases like type 2 diabetes.
A difficulty for Pickrell, though, is that his work relies on datasets that other researchers generate. The community, he said, doesn't always readily share its data. "I've been trying to get people to be more open about sharing published data and things like that, but it's a constant challenge," he said. He added that, ideally, data should be available with "minimal bureaucracy." New mechanisms for sharing data, he said, are needed.
Paper of note
Last year, Pickrell and the University of Chicago's Jonathan Pritchard published in PLOS Genetics a statistical model they developed to examine historical relationships between populations of a species. The goal, Pickrell said, was to develop a tool in which researchers could plug genetic data from a hundred or so populations, and it would spit out how the populations split, mixed back together, and split and mixed again.
The method is based on a series of pairwise comparisons, he added. For example, it calculates how similar French people are to Germans and how similar the French are to the Chinese, and then uses a matrix to reconstruct the history of those populations visually.
Looking ahead
Having access to more and more ancient human DNA is going to change the study of human history, Pickrell said. If you have samples stretching back some 20,000 years, then you don't necessarily need to use statistical methods to try to figure out what was going on at that time; you just turn to the bones.
"That's going to completely rewrite human history in the next decade or so, is what I think," he said. "That's definitely where the field of human history, human adaptation is going."