2PT107 A classification of enzymatic reactions and its application to metabolic pathway search(The 50th Annual Meeting of the Biophysical Society of Japan)

Abstract

TheBiophysicalSociety of Japan General IncorporatedAssociation A huge amoullt molecuLar interaction data are accumu]ated to elucidate entire tandscape ef cellu]ar activities. 3D comp]ex structures of interacting pToteins can add their physical mechanisms, such as interacting residues and dtiving binding forces. The goal of this study is to gather all the reported molecutar interactions, and to predict their 3D structures using hornology modeling as many as poss{b]e, Another goa] is to propose biologically important interacting protein sets to be so]ved their 3D complex structures in the future. However, current experimental molecular interaction data arc not comprehensive and re]iab]e. To overcome these prob]ems, wc eempare interactions of severa] different organisms, and censider conserved interaetions to be highly reliable, As the preiiminary study, we analyzed yeast and human interactromes. The INTACT database stored 70,OOO and 40.000 two-body protein-protein interaetions foT yeast and human, respectively. Ameng these interactions, 3D complex structures for l.5 % yeast and 6 % human pairs can be mode]ed using horno]ogjes for known 3D eomplexes. To reduce false positive data, we extracted conserved interactions (interoiogs) between the two species, Number of the interologs are about 5,OOO, they may be more re]iable and shou]d be biological important. 3D complex struetures can be assigned for 1 1 9S and 24 % of the interlogs for yeast and human, respectively. The remaining 4,OOO protein pairs are good candidates protein pairs to be solved their 3D complex structures. regfim ta77(lbeta=reEptregffmzz{gdimtaes(esem KvlYVtaopHza Deyelopment of re-docking method by obtaining native interaction pattern i, Yuri Matsuzaki2, Masahito Ohue?, Yutaka Akiyamai, (iDept. Ph.vs., Sbh. Sci. & Eng., Chuo Univ, 2Grad. Sck. Sbi., Tbb,o inst. Tbch., ]CIBRC ALS7). it is important to predict which inteTact in living cel] and how these bind each other. We these problems using rigid-body docking algorithrn, generating and ligand 2D structure weTe atso calculated and conslsteney among these three similarlty measures was evaluated.As a resu]t, basieally, pairs with similar protein folds had sjmilar ligands and interactions. However, theTe were many exceptjons, i.e,, dissirnilar inteTactions with similar fotds and ligands, and similar interactions with dissimilar folds and ligands. Novel relationships umong proteins and ligands were mined by our study. 2PTI04Nobuyuki Uchikega Takatsugu Hirokawa3 Comput. &In a study ofprotein-protein Interaction network pairs ef proteins approach to many decoys including false positives. A]though this docking method is popuTar and usefu1, there are some cases with no near-native decoys. In previous work, we developed a method of generated Profile ofInteraction FingerPrints (P-IFP) rather than generating near-native decoys. Then, we applied a profile method to obtaining region including native intcracting residue pairs for re-docking process.We have been developed Interaction FingerPrints (IFP) for the post-...