Infrastructure for Collaborative Protein Structure Prediction

This project had an unusual form: it was designed to promote comparison and exchange of results within the protein structure prediction community.The main activities were centered around web sites and workshops, and the principle challenge was to involve as many of the relevant members of the community as possible.This report covers (1) The CASP structure prediction experiments (CASP2, 1997 and CASP3, 1998 , CASP4 2000, each partially supported by this grant), (2) A community wide experiment to compare potentials for protein structure prediction (termed PROSTAR), and (3) Development of web based systems for the exchange of information on structure prediction methods.In our view, the CASP component of the project has been and continues to be very successful, with participation by almost all members of the protein structure modeling community, and substantial progress in the field to a large extent the result of CASP.This report is written some time after the end of the funding period.Funding ended because DOE abruptly terminated the external computational biology program.Because of CASP's success, it has been possible to continue it with funds from elsewhere, in spite of withdrawal of DOE support, both for this project, and for the Protein Structure Prediction Center, until recently located at Lawrence Livermore.PROSTAR was also successful during the funding period, but has died since, for lack of support.However, much of the functionality and experience was rolled into another resource, 'Decoys are Us', run by Ram Samudrala, formally a graduate student of the Pi's.Two other aspects of the work, comparison of conformational search methods and development of a web exchange system for function assignment to orphan genes, using structure prediction methods, were in progress at the end of the funding period.Function assignment of orphan genes has ceased.Comparison of conformational search has recently been revived, and is the subject of on ongoing experiment in the context of CASP (CASPR), with other funding.Most of this report provides details of the CASP experiments, the most successful and lasting part of the work.