|
A protein is a complex system in itself. Functional sites in protein
structures can be identified by analysis of the protein
residue-interaction graphs, where protein residues are the nodes and their
interactions are the edges. Active-site residues, allosteric paths, ligand
binding sites, and sites sensitive to mutations are all found to have a
high closeness centrality-measure. Conserved sequence regions of protein
families have different compositions of secondary structure elements
(SSEs) from those of the whole proteins. These conserved regions are also
significantly over abundant is some SSE motifs. This can be used to
predict protein functional sites in its own, and in combination with
analysis of the residue-interaction graphs. Residue interactions in
proteins with no known structures can be predicted from multiple sequence
alignments of the protein families using a unique residues pair-to-pair
matrix (P2Pmat) that gives the pair interaction probability for all
160,000 pair substitutions. This matrix can also be used to estimate the
likelihood of predicted protein structures and the nature of changes in
interaction protein residues. More information can be found at: HTTP://WWW.WEIZMANN.AC.IL/SARIG HTTP://WWW.CCBB.PITT.EDU/p2p http://bioinfo.weizmann.ac.il/funcwizer/ |
|