Trieste.pptVIP

Motif Discovery Wei-Mou Zheng Inst Theor Phys, Academia Sinica zheng@itp.ac.cn Two motifs in 5 sequences Outline Protein structure A brief review of secondary structure prediction Hidden Markov model: simple-minded Hidden Markov model: realistic Discussion References Discussions HMM using refined conformational states and window scores is efficient for protein secondary structure prediction. Better score system should cover more correlation between conformation and sequence. Combining homologous information will improve the prediction accuracy. From secondary structure to 3D structure (structure codes: discretized 3D conformational states) Discrete conformational states Secondary structure states: helix + strand ~30+22% Loop Single secondary structure may vary significantly in 3D. Phase space partition: Ramachandran plot (Rooman, Kocher Wodak, 1991) Library of representative fragments (Park Levitt, 1995) representative points in the phase space General clustering (Oligons; Micheletti et al, 2000) Hidden Markov models (Camproux et al., 1999) distributions as modes pros: connectivity effect (correlation) cons: many parameters, inconvenient for assigning states to short segments Mixture model: clustering based on peaks in probability distribution y j 0 0 p p -p -p E B N T H I F P M L U G E O * S E: extended B: beta I,H: alpha L,M: left T,U: turn P: Pro-rich G: Gly-rich N: NH-NH conflict O: CO-CO conflict S: side chain conflict *: NH-CO conflict AN APPROACH TO DETECTION OF PROTEIN STRUCTURAL MOTIFS USING AN ENCODING SCHEME OF BACKBONE CONFORMATIONS H. Matsuda, F. Taniguchi, A. Hashimoto The placement of a protein backbone chain with an icosahedron in the Cartesian coordinates. Representation of 3D backbone: C-alpha pseudobond angles Four-residue fundamental unit: btb r1r2r3r4 1 2 3 4 b2t3b3 b2 b3 t3 ~ 3.8 A S1 S2 Sliding windows Pseudobonds: bend and torsion angles dihedral angle, si