WU team places 4th in scientific challenge

| Staff Reporter

The “Phe-Thr-Trp” isn’t just a bunch of random sounds strung together. Rather, it is the name of a team from Washington University that placed 4th out of 25 teams in the 2010 University Protein Folding Challenge.

This challenge, composed of students from top-ranking universities, was hosted by Foldit, a computer application developed by the University of Washington that enables people to contribute to scientific research in collaboration with computers. MedImmune, an international biologics business that aims to create drugs to improve health, sponsored the competition.

This contest, which began Friday, Nov. 5, challenged the teams from universities such as CalTech, M.I.T. and Stanford with the structure of an over-expressed protein linked with the expression of pancreatic cancer. The goal of the competition was to use the Foldit software to fold the protein to optimal structure, essentially curing cancer. The team that returns the protein to a structure closest to its optimal structure wins the competition.

The challenge ended Thursday night, and the winner, “Team Crystallin” of M.I.T., received a prize of $5,000. The Washington University team, “Phe-Thr-Trp,” was composed of graduate students hailing from the departments of immunology, genetics and computational biology—Ryan Sullivan, Kevin Forsberg, Igor Nikolskiy and Subhajit Poddar—and ranked 4th in the competition.

This competition has significant implications beyond being a competitive game.

Proteins are very important to human life and knowledge: They control everything that goes on in one’s body, from enabling blood flow to sending signals between the brain and rest of the body. Proteins are composed of amino acids, which create a unique shape for the protein. Sometimes these shapes are warped from what their forms should be. As the shape of the protein determines the protein’s function, misshapen proteins are often the cause of many biological problems. Foldit uses humans’ puzzle solving abilities to predict the structure of a protein.

“It is difficult for computers to fold proteins,” Sullivan said. “We have these massive super computers folding proteins all day, but there are certain things that the human mind can do better. The goal of the game [Foldit] is to see how humans would solve the problem, what do they do, so you can teach the computer to think like a human.”

Essentially, the participants are sharing intuition and logic with computer software. Computers and humans both have unique advantages as well as disadvantages, and this competition forges a connection between the two in the pursuit of more efficient knowledge.

Given enough time, a computer could find the optimal structure of any protein, but this process is helped by human intuition.

The process of the competition proves itself more significant than the competition’s final stage. Foldit holds huge implications for the future of drug design. Problems with proteins are speculated to be the main factor causing diseases such as Alzheimer’s, HIV/AIDS and cancer. Humans’ use of Foldit enhances scientific knowledge about the ways proteins fold. This can allow drug companies, such as MedImmune, to learn more about the proteins and thus know how to better target these proteins with drugs.

Drug companies normally test numerous random drugs because they do not know what the targeted protein looks like, and this costs billions of dollars. With Foldit and human collaboration, drugs can be tailored more specifically to the known shapes of proteins, saving time and money.

Overall, the competition was a success. Cash prizes were awarded to the top three teams in the competition, and the Washington University team established themselves as a top competitor among high-ranking universities. This is just the beginning, as major scientific achievements in regard to proteins and diseases are bound to spring from this revolutionary method.

Sign up for the email edition

Stay up to date with everything happening at Washington University and beyond.

Subscribe