Automatic cDNA Classi cation System for Mouse Genome Project

Classi cation of cDNA/protein sequence data is necessary for analysis and further application of sequence data, i.e. understanding of pathway of proteins in a cell and signal transduction. Since rapid increase of the number of cDNA/protein sequences is expected, the procedure of the classi cation must be systematic and automated to avoid human eye inspection. In this study, the sequence data are classi ed based on two terms; subcellular location and function. Again, false positive/false negative results are always expected in the use of prediction method, and the human inspection is necessary to avoid this problem. We proposed a method to reduce the false positive for PROSITE motif search by using a structural information.