Blocks-based methods for detecting protein homology

The most highly conserved regions of proteins can be represented as blocks of aligned sequence segments, typically with multiple blocks for a given pr otein family, The Blocks Database World Wide Web (http://blocks;fhcrc.org) and e-mail (blocks @ blocks. fhcrc.org) servers provide tools to search DNA and protein queries against the Blocks+ Database of multiple alignments. W e describe features for detection of distant relationships using blocks. Bl ocks+ includes protein families from the PROSITE, Prints, Pfam-A, ProDom an d Demo databases. Other features include searching Blocks+ with the BLIMPS and NCBl's IMPALA programs, sequence logos, phylogenetic trees, three-dimen sional display of blocks on PDB structures, and a polymerase chain reaction (PCR) primer design strategy based on blocks.