Constrained Molecular Dynamics Simulations

Period of Performance: 08/01/1998 - 01/31/2000


Phase 1 SBIR

Recipient Firm

Virtual Chemistry, Inc.
Redwood City, CA 94065
Principal Investigator


This proposal is concerned with the development and commercialization of a novel algorithm for propagating constrained equations of motion (EOM). The objective is to provide a robust and general solution to this complex problem. Solving the constrained EOM in chemistry allows one to remove uninteresting degrees of freedom (e.g., bond vibrations) from the system and speed up the simulation. The algorithm we propose to commercialize, called the constraint-force (CF) algorithm, was formulated at the Jet Propulsion Laboratory (JPL) for robotics systems. The CF algorithm promises to be the most efficient and general means of propagating constrained EOM in MD simulations. The CF algorithm can be incorporated into the MD software provided by all major computational chemistry software companies today. The routine availability of the CF algorithm will allow the creation of strategies for conformational searching and rapid energy equilibration. Such tools will benefit medicinal chemists as well as traditional molecular modelers, thus leading to a much needed expansion in the computational chemistry market. The CF algorithm also has the interesting advantage that it is inherently parallel: Not only will the algorithm be the most efficient for a single processor, it can also scale up to large numbers of processors. This is important for researchers on the cutting edge of simulation science doing large-scale/long-time MD simulations of million atom systems. PROPOSED COMMERCIAL APPLICATIONS: Virtual Chemistry, Inc. intends to license this technology to software vendors, academic sites, and industrial researchers with active molecular dynamics programs. A key to making this successful is the implementation of a portable CF EOM package along with consulting expertise from Virtual Chemistry for assistance with integrating the technology with existing efforts. The ability to simulate the relevant dynamics of significant (large) biological ensembles and to rapidly explore drug-receptor conformations will have important uses in health related research.