Ab initio calculations of GTP hydrolysis by using NWChem and TAO components
EMSL Project ID
3431
Abstract
As part of the Common Component Architecture (CCA) project funded by the U. S. Dept. of Energy (DOE) under the Scientific Discovery through Advanced Computing (SciDAC) program, we are developing NWChem components and evaluating the performance of these components within a CCA compliant framework, ccaffiene (http://www.cca-forum.org/ccafe/).The purpose of the project is to introduce new algorithms and capabilities into NWChem that are present in other components developed in DOE national laboratories. To avoid duplication of effort required for re-implementation, the CCA approach is utilized (http://www.cca-forum.org). In this approach, the data/functionality is defined through shared interfaces and the framework enables the components to be ?connected? through these interfaces. The involved components are
1. NWChem (http://www.emsl.pnl.gov/pub/docs/nwchem) and MPQC (http://aros.ca.sandia.gov/~cljanss/mpqc/) for energies, gradients and Hessians
2. Optimization computations within TAO (http://www-fp.mcs.anl.gov/tao/),
3. Linear algebra computations within Global Arrays (http://www.emsl.pnl.gov:2080/docs/global/ga.html) and PETSc (http://www-fp.mcs.anl.gov/petsc/index.html)
The present project focuses on investigating the performance of different optimization algorithms in NWChem and TAO. To test performance we will apply different optimization algorithms for studying the Potential Energy Surface (PES) at an ab initio level for biologically important chemical compounds. For each reaction described below, we will need to perform several optimizations to compare the performance of NWChem compared to those of the TAO optimization component. As new TAO components and optimization methods become available, we will also utilize them for studying biologically important reactions.
GTP-binding proteins (GTPases) play an important role in a variety of biological processes such as cell growth and transcription control. Particularly, the Ras family proteins of GTPases play an important role in linking receptors on the plasma membrane to signaling pathways. Ras mutants are found in 25-30% of human tumors. When GTPases are bound to GTP these proteins are active. Later, when GTP is hydrolyzed to GDP, these proteins become inactive:
Ras-GTP + H2O -> Ras-GDP + HPO42-
The hydrolysis can be catalyzed by GTPase-activating proteins (GAPs). In spite of the obvious importance of GTP hydrolysis, the mechanism is still controversial. To explore various GTP hydrolysis mechanisms, we will compute acidity constants of GDP, GTP, Ras-GTP, and other related chemical compounds as an initial step. In the following step the role of GAPs in hydrolysis will be examined. The preliminary computations suggest that geometry optimizations in gas-phase and water solution, which are needed to compute acidity constants, will be time extensive. To complete such a large number of calculations will require the maximum 75,000 hours of computation time on the HP Supercluster.
Project Details
Project type
Capability Research
Start Date
2003-03-21
End Date
2004-03-23
Status
Closed
Released Data Link
Team
Principal Investigator
Team Members