This subproject is one of many research subprojects utilizing theresources provided by a Center grant funded by NIH/NCRR. The subproject andinvestigator (PI) may have received primary funding from another NIH source,and thus could be represented in other CRISP entries. The institution listed isfor the Center, which is not necessarily the institution for the investigator.The identification of protein homologs was traditionally based on local sequence similarities with data base search methods such as PSI-BLAST and FASTA. Recent methods which were designed to identify more remote homologies extended their predecessors by exploiting global features of the topology induced by the sequence similarity relations, or by using additional data sources, such as protein protein interactions (PPI).In this project we use two of these recent methods to search for homologous structures of the kinetochore of Saccharomyces cerevisiea in Schizosaccharomyces pombe and Homo sapiens. Specifically, our analysis is based on the following two methods: (i)We use the machine learning technique of network propagation and adjust the rank propagation (RANKPROP )algorithm to handle a hybrid network composed of both PPI and sequence similarity links. Propagating in the hybrid network yields a list of candidate proteins, predicted to be related to the S.cerevisiae kinetochore.(ii)PPI network alignment using the Graph BLAST algorithm. The generation of global alignment between two PPI networks is based on the pairwise sequence similarities of their proteins and identifies dense clusters of conserved interactions. Aligning the Kinetochore subnetwork of S.cerevisiae with the PPI networks of H.sapiens and S.pombe allows us to detect candidate subnetworks in these species which are predicted to be related to the S.cerevisiae kinetochore.
Showing the most recent 10 out of 583 publications
