Bench Press

Berkeley researchers are on the heels of settling debates over taxonomy using innovative computing methods.

Apparently, whoever came up with the saying “nothing good ever comes from cheating” didn’t go to Berkeley. Recently, a team of researchers at the University of California, Berkeley took a little bit of inspiration from plagiarism-detecting software and created a program which can compare the entire genomes of two distinct organisms. Using what are known as feature frequency profiles (FFP), professor of chemistry Sung-Hou Kim and his team were able to successfully compare the genomes of several different organisms, and in a sense, “map” the organisms’ evolutionary family tree.

Traditionally, methods of determining how closely related two species are focus on a very specific subset of genes that the organisms have in common. The differences and similarities in the genetic code are then counted up and a computer program constructs the family tree. The more differences in the genes, the more distantly related the organisms are. However, the drawback of this technique is that it relies on organisms having these specific genes in common. What may end up happening is that an organism may not have a “homologous” gene to compare it with. Additionally, two genes that are used for comparison oftentimes conflict with each other; one gene says two organisms should be closely related while another one says they shouldn’t. With this innovative new approach of using FFP’s, the entire genome is sequenced and compared, as opposed to several different genes, allowing scientists to view differences in a larger scope.

Maybe even more surprising, this idea of using FFP’s as a means of comparison transcends the realm of genetics. When applied to literary works, this algorithm was able to better detect similarities between texts by the same author, of the same genre, and of the same historical era, than other conventional methods.

I was just stunned when I saw this,” Kim said. One of the reasons this method works better, he said, may be that, while word frequency analysis treats each word independently, feature frequency analysis picks up syntax.

Armed with this technique, these researchers have successfully segregated the proteomes of bacteria, Archaea, and eukaryotes with genomes of varying complexity into distinct groups and domains. Whenever disagreement occurred between their findings and common scientific belief, there was generally some kind of ongoing debate over that particular organism’s taxonomy. Professor Kim and the rest of his colleagues hope that they can use this algorithm to classify some enigmatic viral genomes and possibly utilize it in other fronts, such as literature and electronic encoding.

To me, this breakthrough in comparative genomics is a chance to reflect on how marvelous and miraculous life and evolution are. To think that the entire human civilization arose from single celled organisms, and that the building blocks and impulses which act upon bacteria and micro-organisms to keep them alive are the very same mechanisms that we live on.