Uwe Ohler: The promoter coder
Assistant Professor, Computational Biology. Age: 37
Uwe Ohler spent a year in Berlin serving meals at a homeless shelter after finishing his undergraduate degree in computer science at the University of Erlangen–Nuremberg. It was the mid-1990s, the Human Genome Project was well underway, and Ohler felt like he was missing out on something. He always took an applied approach to computer science, and he had dabbled in genetics in college, applying speech recognition algorithms to find new bits of regulatory sequence in DNA. But now he realized that he wanted to study computational biology to fully explore the intricacies of the human genome. After getting off work at the shelter in the evenings, he would walk next door to a Technical University of Berlin computer lab to fill out applications for graduate funding.
METHODS: Starting in 1998, as a PhD student, Ohler worked with data generated in the Drosophila Genome Project, splitting his time between the University of Erlangen–Nuremberg and the University of California, Berkeley. As more data accumulated, co-adviser Gerry Rubin asked him to stay in Berkeley after graduation. He spent 9 months there using RNA sequences to pinpoint genes. Then he probed the surrounding DNA for promoters, sites that turn on gene expression.
“He was a delight to have around—quiet but highly effective,” writes Rubin, who is now vice president and director of Howard Hughes Medical Institute’s Janelia Farm Research Campus, in an email. Researchers had found only a handful of promoter sequences in the fruit fly genome by 1998, but when his group found several more, “we were absolutely surprised,” Ohler says. From those results Ohler created a tool, dubbed McPromoter, to help others improve promoter recognition.1
RESULTS: Ohler has increasingly pursued biological questions in his research. As a postdoc with Chris Burge’s group at the Massachusetts Institute of Technology, Ohler studied how human and mouse sequences produce slightly different forms of mRNA, and thus protein. Using an algorithm Ohler devised, the group estimated that several hundred pieces of coding DNA in the mouse and human are actually spliced out before they are translated in proteins.2 This work pointed to new variations or forms of genes that had been missed, and helped explain how new protein isoforms arise.
DISCUSSION: Now at the Duke Institute for Genome Sciences & Policy, Ohler and his group are designing new algorithms, based on high-throughput sequence datasets they generate, to find promoters. But he’s also entered a completely new arena: image analysis. He’s developed a way to acquire and analyze microscope images of Arabidopsis, and using the software, his group can depict relative gene expression for different cell types in the plant’s roots.3
Ohler brought unique skills to the table when bioinformaticians were struggling with genetic data. “In the field of computational biology, there are a lot of computer scientists who come in and dabble but they don’t make a practical contribution,” says Burge. “But Uwe developed something really useful. That’s why he stood out.”
1. U. Ohler, H. Niemann, “Identification and analysis of eukaryotic promoters: recent computational approaches,” Trends Genet 17:56–60, 2001. (Cited in 102 papers)
2. U. Ohler et al. “Recognition of unknown conserved alternatively spliced exons,” PLoS Comput Biol 1:113–22, 2005. (Cited in 30 papers)
3. D.L. Mace et al. “Quantification of transcription factor expression from Arabidopsis images,” Bioinformatics 22:E323–E331, 2006. (Cited in 7 papers)