Skip redundant pieces
KU Home  :  KU News

KU News Release

More Information


University Relations

p (785) 864-3256
f (785) 864-3339
Nov. 21, 2006
Contact: Brandis Griffith, University Relations, (785) 864-8855.

KU researcher wins international protein prediction competition by wide margin

Yang Zhang defeated opponents by a wide margin in the Critical Assessment of Structure Prediction competition. His computer server was able to take protein sequences through a process that generates their structure. His model came closer than any other competitor to the actual structure. Being able to generate protein structures quickly will aid in drug discovery and design.

LAWRENCE — A University of Kansas researcher has beaten opponents by a wide margin in a competition that could lead to new drug treatments for various illnesses.

Yang Zhang, assistant professor of molecular biosciences, won first place in the prestigious automated server predictions category at the Critical Assessment of Structure Prediction competition.

Protein structures are 3-D illustrations of the amino acid sequences that make up a protein, or polypeptide chain, and give it its specific biological function.

“In order to know the function, we must know the structure, because the structure decides the function,” Zhang said. “The final goal is we can design new drugs.”

Finding the structure of a protein is a long and expensive process, said Zhang, also a faculty member at KU’s Center for Bioinformatics. To identify one protein can cost upwards of $1 million, and there is an infinite amount of proteins in nature.

The international competition is “blind” because the protein structures are new, unpublished and therefore unknown.

To win the competition, which takes place biannually, Zhang’s lab wrote a computer software program. The computer server generated the protein structure within the 48-hour time limit. Zhang’s server is called the I-TASSER — Iterative Threading Assembly Refinement. It is a newer version of TASSER, which Zhang developed with Jeffrey Skolnick, a professor at the Georgia Institute of Technology.

“If we have good software to predict the structure from the sequence, we can generate as many structures as we want in a very quick, easy manner,” he said. “That is our goal.”

Zhang competed against 98 others in the server competition.

The winners are selected by measuring their predictions against a structure created by a control group. Zhang beat the second place winner by 5 percent to 10 percent.

“I’m happy that our method really worked well,” he said. “Drug companies pay attention to the competition because if you win it, companies want to buy the software because these high-resolution protein models are very important to drug design.”

This is the seventh year for the competition. Zhang has been invited to speak at the Critical Assessment of Structure Prediction conference in California, which will be held Nov. 26-30.

Zhang has just completed his first year on the faculty at KU. Before coming to KU, he worked with Skolnick at the University at Buffalo.


The University of Kansas is a major comprehensive research and teaching university. University Relations is the central public relations office for KU's Lawrence campus. | (785) 864-3256 | 1314 Jayhawk Blvd., Lawrence, KS 66045