Dennis Wall: From moss to autism
In 1996 a few hundred plant scientists gathered in Baton Rouge, La., for an annual phylogenetics meeting. Biology undergraduate Dennis Wall rushed into a lecture hall to meet Brent Mishler, a University of California, Berkeley, integrative biologist, who was considering taking on Wall as a doctoral student. But when Wall, late and disheveled, tried to climb over a row of folding chairs to get a closer seat, he tripped, and the chairs crumbled beneath him. When he got up, applause filled the auditorium.
Researchers including Mishler, who became Wall's PhD advisor, say Wall would continue to surprise them—academically, that is—for years to come.
As a PhD student, Wall tracked the radiation of an obscure genus of moss from its origin in Malaysia to its outermost ranges in French Polynesia. He wanted to see how quickly branches...
Wall applied the epidemiological model to genetic data and showed that the moss diversified at a blistering evolutionary pace.1
As a postdoc in biologist Marcus Feldman's Stanford University lab, Wall analyzed more than 3,000 proteins in four yeast species, and showed that both the expression and "dispensability" of a protein can affect the rate of its evolution, something that researchers suspected but few had convincingly shown.2
In 2003, Wall became the director of Harvard Medical School's Computational Biology Initiative. There he created an online content management system that integrated tools used by many HMS researchers to analyze DNA and amino acid sequences. In April 2006, Wall took a faculty position at the newly formed Center for Biomedical Informatics at HMS. He faced a number of challenges, says Isaac Kohane, associate professor of pediatrics who recruited Wall and formed the center. Wall wasn't trained in medicine. He had to figure out which questions he could answer using phylogenetic techniques, and he had to harvest the right data for these meta-analyses. "I have to admit that I was silently skeptical," Kohane says. "Boy, was I wrong."
Drawn to autism in part for its biological complexity, Wall and his group began mining the literature and integrating disease determinants—genes, transcription factors, and proteins—that could help generate a network of autism gene candidates.