Protein arayüzlerinin kümelenmesi ve bunların analizi

Improvements in experimental techniques increasingly provide structural data relating to protein-protein interactions and across-interface interacting domains. Classification of structural details of protein-protein and inter-chain domain-domain interactions can provide valuable insights for modeling and abstracting design principles of protein-protein interactions. This dissertation mainly focuses on clustering protein-protein interactions by their interface structures, exploiting these clusters to obtain and analyze shared and distinct protein binding sites, and dissecting the domaindomain combinations in protein interface structures. 22604 unique interface structures in the PDB are found and these unique interface structures are used to find protein pairs which have multiple binding sites to interact with each other. For the first time, domain combinations in the interfaces both on the same and on the complementary monomer are analyzed, based on PFAM domains classification. The domain combinations are represented by constructing a network. In addition to PFAM domain analysis, fold, family and superfamily classification of SCOP are mapped to unique protein interface structures in order to find the SCOP domains in each classification level which are the mostly frequent ones used in protein interfaces and have the most different unique interface structures. This comprehensive study of protein interface structures and interface domain combinations can serve as a resource for the community.