By Joshua Baxt, CSE writer
UC San Diego Computer Science and Engineering professor Vineet Bafna has been named a 2019 ISCB Fellow by the International Society for Computational Biology. ISCB fellowships honor members who have made outstanding contributions to computational biology and bioinformatics.
Bafna is being honored for his efforts to leverage bioinformatics to better understand cancer and other traits. He has focused on three aspects of molecular biology: genomics, population genetics and proteomics.
Cancer Genomics
Cancer genomes are chaotic, by definition. While genomic sequencing has great potential to improve patient care, researchers and clinicians need better tools to interpret the data. Copy number variations – in which oncogenes get duplicated, amplifying their signals – are common mutations. But it can be difficult to measure their impact.
“There’s been a lot of research to suggest that when you have high copies of oncogenes, or genes that are proliferative in nature, that leads to poor outcomes,” says Bafna. “But what is not clear is how those copy numbers variations arise. How is it that a gene can make 20, 50, 100 copies?”
Bafna and colleagues worked on this problem for some time before finding the answer. In a paper published in Nature in 2017, Bafna, UC San Diego Pathology Professor Paul Mischel, and colleagues, determined that large DNA sequences could go rogue and duplicate oncogenes.
“We found the increase in copy numbers happen because of ‘little’ pieces of DNA that break away from the chromosomes, circularize and replicate, making many copies of those pieces, which are then passed on to the daughter cells,” says Bafna. “Millions of base pairs long, they often contain oncogenes, massively amplifying the oncogenic signal. People didn’t realize extrachromosomal DNA could play such a big role in amplification.”
Graduate student Viraj Deshpande played a big role in this work, developing software that reconstructs the fine structure of these extrachromosomal DNA elements.
How Populations Adapt
Our ancestors first evolved pretty close to sea level. As humans migrated, they sometimes took to the mountains, becoming accustomed to reduced oxygen levels. But which genes drove those adaptations?
Working hand-in-hand with cell biologists, the Bafna lab analyzed genome data from Andean populations, looking for mutation frequency patterns that reveal genetic adaptations. They identified a few large genomic regions, which lead them to novel genes that helped humans adapt. One problem remained: there could be as many as 50,000 mutations in that region, hitchhiking with the few mutations driving the adaptation. It was the classic needle in a haystack.
“That problem was widely considered unsolvable, with only one relevant paper appearing in 2010 in Science,” says Bafna.
But through a combination of population genetics and machine learning, the team figured out which genes were mediating the adaptation to altitude, as well as outlining the genetic mechanisms behind chronic mountain sickness.
Proteogenomics
Molecular biology is a relatively young science. In other words, there are still big gaps in our gene and protein knowledge.
“Our research is based on the premise that the set of proteins produced by the genomics community is incomplete,” says Bafna.
Using the genome as a template, the lab has been using mass spectrometry to identify proteins and peptides (pieces of proteins) mutated in cancer samples. These methods have identified new genes and protein variations that may be implicated in cancer.
Through these and other efforts, Bafna’s collective work has made an enormous impact on both computational biology and cancer research.
“ICSB is the premier society for computational biologists.” says Bafna. “I am honored to be selected as a Fellow.”