We propose to facilitate the understanding of protein functions and metabolic processes in living species, by developing a computer system for completely automated prediction of biochemical or biological functions for large sets of protein sequences. Instead of the conventional approaches that typically rely on one round of database search and inherit the annotation from the best-scoring sequence match, we will construct a proprietary structured database of complete proteomes and develop an automated multistep strategy of database searches, feature prediction, and annotation. The cascade of analysis will include: selection of the reference databases, flexible filtering, domain dissection, iterative searches, motif analysis, and rule-based modification of the database annotations. The system to support this approach will be capable of analyzing one protein every five minutes, which is approximately a tenfold increase over the productivity of a trained analyst. In addition, custom heuristic algorithms will result in 20-30 percent greater accuracy in annotations compared to the existing automated approaches. The rapid and accurate prediction of protein functions at the industrial scale established in the proposed project will be applied for exhaustive annotation of newly sequenced proteomes and EST collections, in order to reconstruct in detail biochemical pathways targeted for modification in various organisms.
The development of the genome-scale functional annotation of protein sequences at high speed and accuracy will accelerate gene discovery, target validation and metabolism reconstruction. This will provide for rapid and reliable development of human therapeutic molecules, antimicrobials, agricultural products, and industrial enzymes. Many new functional predictions for proteins from different species will be made and the value-added sequence databases will become available.
