DIMACS Workshop on Sequence, Structure and Systems Approaches to Predict Protein Function

Presented under the auspices of the DIMACS/BioMaPS/MB Center Special Focus on Information Processing in Biology.
*********************************************************************
While more than two hundred complete genomes have been sequenced, a large fraction of genes in genomes do not have assigned functions, and elucidation of the biological roles of all unknown gene products remains an elusive goal. In addition to genomic sequence data, in recent years we have witnessed an explosion of structural data, along with large-scale protein network data. This has led to a number of complementary approaches to predict protein function based on heterogeneous data from diverse experimental sources. The goal of such methods is to decrease the amount of time-consuming experimental work necessary to interpret the complexity of genomes and proteomes. In general, protein function can be predicted by the analysis of specific conserved structural and sequence features, by transferring the annotation from experimentally characterized genes to their uncharacterized homologs , by genome context and cross-genomic analysis, and by analyzing proteins within the context of biological networks. Each of these methods faces unique computational and statistical challenges.
The goal of the workshop is to bring together biologists, computer scientists and mathematicians who work on various aspects of protein function prediction. This workshop will provide both a venue for reviewing the current state-of-the-art of diverse methods as well as a platform for further cross-fertilization and integration of sequence, structure and systems approaches.