A Protein Class Database Organized with ProSite Protein Groups and PIR Superfamilies

Abstract

A protein class (ProClass) database is developed as a "value-added" "second-generation" database organized according to family relationships. The database collects non-redundant protein sequence entries from SwissProt and PIR databases, and classifies them in families defined collectively by the ProSite protein groups and PIR superfamilies. The major objectives of the database are to maximize family information retrieval, to provide speedy family identification, and to help organizing existing protein sequence databases. The database has two sub-databases: PCFam (ProClass Family) to define protein families and provide links to ProSite patterns and PIR superfamilies, and PCSeq (ProClass Sequence) to describe sequence entries and provide links to PCFam, SwissProt, PIR, and ProSite databases. The current ProClass release has a total of 85,165 sequence entries, about half of which are classified in 3072 ProClass families; it also contains 10,431 newly established SwissProt-PIR links. The database can help reveal domain structures of related families, define new ProSite and PIR families, and provide family assignments for unclassified sequence entries. New ProSite and PIR family members are readily identified via database cross-reference, including 9437 SwissProt entries and 8522 PIR entries. False negative family members missed by both ProSite and PIR are detected using a neural network family identification system. The newly identified superfamily memberships are being incorporated into the current PIR database releases in a collaborative effort with the PIR. The ProClass database is accessible through anonymous FTP and on-line search on the World Wide Web.

Artificial Intelligence in Biological Data

“…Then recalculate the distance matrix using average distance between the newly constructed node and the remaining leaves. The calculation is done iteratively until only two clusters of nodes are left [86][87][88].…”

Rational Design, Selection and Specificity of Artificial Transcription Factors (ATFs): The Influence of Chromatin in Target Gene Regulation

“…For a highly specific TF, such as a 6ZF ATF, it is expected to have many fewer binding events in the genome. The integration of genome-wide ChIP-on-chip experiments [62][63][64][65] and bioinformatics approaches (such as ProSITE [66][67] using the predicted ATF-binding site as a probe) are perhaps the most accurate way to evaluate the genomic specificity of a given ATF. However, genome-tiled arrays required for genomewide ChiP-on-chip assays are fairly expensive and involve specialized technology, perhaps explaining why these experiments have not extensively been used for the study of in vivo ATF-DNA interactions.…”

ProClass protein family database

“…Family classification requires two crucial elements: search tools that fully utilize the information embedded within families of homologous sequences, and databases organized according to family relationships. We have developed an integrated database and search system, consisting of the ProClass protein family database (1) and the GeneFIND (Gene Family Identification Network Design) search program (2).…”