• Subscribe

Feature - PKU protein structure

Feature - Peking University advances protein research


At top: Superimposition of 20 structures of hPHPT1. At bottom: Ribbon diagram of the mean structure of hPHPT1. Red: alpha-helix, blue: beta-sheet, white: random coil. Image courtesy of Weibin Gong

Researchers at Peking University's Beijing Nuclear Magnetic Resonance Center (BNMRC) can determine 3D protein structures 1000 times faster, thanks to the grid.

Function follows form in the protein world, so their work contributes to scientists' understanding of the roles that individual proteins play. This in turn should lead to advances in drug discovery and health care.

The Peking University team uses Nuclear magnetic resonance (NMR) spectroscopy to determine protein structures. NMR works well, but the calculations involved are time-consuming and computationally demanding. Scientists have so far learned the structure of only about 50,000 proteins out of the millions that exist in nature.

The Peking team has calculated the structure of the human protein PHPT1, important in regulating the transfer of signals in certain cell types.

The team starts work on a new protein by determining experimentally its structural constraints. Then they run a set of simulations to determine a likely structure for it, optimizing for stablity. The more stable the structure, the more likely a real protein is to assume it.

Using grid resources through EUChinaGRID, a project supporting the interconnection and interoperability of existing European and Chinese grid infrastructures, the team runs the Amber molecular dynamics software package. Amber first calculates a set of possible protein structures. Then, through a process called "simulated annealing," the software simulates heating each calculated structure to around 2,000 degrees Kelvin and cooling it slowly to absolute zero. By doing this, the simulated proteins tend to "freeze" into the most stable, i.e. the lowest energy, structures possible.

At the end of this calculation-simulation process, the researchers average the most stable 20 percent of the results, adjust the constraint parameters according to this intermediate result, and run the process again. They typically run it about ten times in succession, at which point they consider their simulated structure to be sufficiently stable.

"Without the grid, these calculations would be extremely time-consuming," said Peking University researcher Weibin Gong.

Each round of calculations for a particular protein requires the computation of 200 structures, and each structure takes about four hours to compute on an Intel 2.4 GHz CPU, said Gong. Each protein requires about ten rounds of calculations, so it would take around 8,000 hours (almost a year) to produce one final protein structure on a single processor. In contrast, by using the BEIJING-PKU work-node of EUChinaGRID, the calculation for the PHPT1 protein took just over four hours to complete.

"NMR protein structure calculations require extensive computing resources," Gong said. "The grid allows us to access several CPUs and distribute our calculation tasks, thus greatly accelerating our protein structure calculations."

-Amelia Williamson, for iSGTW

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2018 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.

Republish

We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.