UCSC BME 220 Spring 2009 schedule

UCSC BME 220 Schedule

Protein Bioinformatics, Spring 2009

(Last Update: 12:27 PDT, 5 June 2009 )

Date	Topic
30 Mar 2009	Administrivia, project list
1 Apr	more project list
3 Apr	Overview of template-based prediction, difference between query-based approach and high-throughput approaches like ModBase, different meanings of "domain" (independently folding, compact structure, evolutionary conservation), psiblast as example of iterated search
6 Apr	Discussion of CASP, different prediction categories, how to read distance vs. % of residues assessment plots
8 Apr	Example of SAM-T08 protocol for "easy" target: T0396
10 Apr	Proposal due; Discussion of hw1 (using SAM-T08 protocol on beta-defensins); example of SAM-T08 protocol for hard target:T0476; consensus metaservers (median GDT_TS)
13 Apr	Local structure alphabets: DSSP, Dssp-ehl2, Stride, str, str2, Bystroff
15 Apr	Local structure alphabets: deBrevern's protein blocks, str4, alpha, n_sep, o_sep, n_notor, o_notor, n_notor2, o_notor2, CB-burial-14-7, near-backbone-11, CB8-sep9; intro to neural nets: perceptrons, 2-layer Hopfield feed-forward networks, PhD and PSIPred 2-network approach
17 Apr	HW 1 due Review of 2-layer networks. Multi-layer networks, softmax, log loss, windowing, back-propagation, RPROP. Multiple-start training protocol.
20 Apr	RDB and gnuplot for handling and viewing data. The version of RDB we use (based on version 2.6d) can be found in /projects/compbio/bin/scripts. The official version was at ftp://rand.org/pub/RDB/hobbs/ but that link no longer seems to work. There appears to be a version 2.6d on the FreeBSD mirror sites
22 Apr	Started explanation of undertaker commands,going through first part of try1.under, costfcn-init.under, and try1.costfcn
24 Apr	Draft 1 due Went through predictions for HW1, looking at what went wrong. Mainly path set incorrectly or sequence-target name mismatch.
27 Apr	undertaker cost functions: align_constraints, ehl2_constraints, alpha, alpha_prev, break, burial cost functions, radius of gyration, phobic_fit, knot
29 Apr	undertaker cost functions: sidechain, n_ca_c, bad_peptide, hbond_geom (and other hbond cost functions)
1 May	HW2 delayed until Monday (no one had it ready). Other use for local-structure neural-nets: Josue's short-protein predictor and undertaker cost functions. Firas's knot-finder algorithm. Short digression on http://Fold.It
4 May	HW2 due Undertaker "TryAllAlign" and "OptConform" commands
6 May	Bill Scott, X-ray crystallograoher (Kevin absent, University College Dublin)
8 May	Bill Scott, X-ray crystallograoher (Kevin absent, University College Dublin)
11 May	Draft2 due Overview of model building: conformation generation, cost function, and stochastic search. Different styles of stochastic search: simulated annealing, replica-exchange Monte Carlo, molecular dynamics, genetic algorithms.
13 May	Conformation generation: initial models and conformation change. Rosetta and NMR: initial fully extended chain. Rosetta: initial model from alignment, with loop library. Undertaker: initial random model, with alignment insertion. Fragment replacement in Rosetta and undertaker. Undertaker conformation change operators: InsertFragment, InsertSpecificFragment, InsertAlignment (and discussion of tree structure used for non-contiguous chain) InsertSSBond, TwoFragment, ReduceBreak, ReduceClash, ReduceConstraint
15 May	The rest of undertaker's conformation-change operators. tweaks: TweakPhiSubtree, TweakPhiSegment, TweakPsiSubtree, TweakPsiSegment, TweakPsiPhiSubtree, TweakPsiPhiSegment, TweakHbondSubtree, TweakHbondSegment, TweakMultimer, TweakPeptide, ImproveSSBond, FixOmega rigid transform of subpart: JiggleSubtree, JiggleSegment, OptSubtree, OptSegment, OptAllSegments, ShiftSubtree, ShiftSegment Gap closing: ReduceBreak (done last time), CloseGap, MoveGap, HealGap (also HealPeptide, though that is not really gap-closing, since it is the same operation as HealGap, but applied to a peptide plane that is not identified as a gap) Crossover operators: CrossOver, DoubleCross, CrossAndInsert Backrub and BigBackrub Sidechain operators: OneRotamer, ClashingRotamer, ClusteredRotamer
18 May	Kevin absent (CA state science fair) Jenny Draper (lab tour: protein expression and purification)
20 May	Kevin absent (CA state science fair) Dietlind Gerloff: protein-protein docking ---------------------------------------------
22 May	Draft3 due, Also select journal-club papers Poster design. See also https://intranet.soe.ucsc.edu/lab-support#PosterPrinting Discussion of mutal information and predicting local structure from single residue (or designing residue from single local structure label). Best alphabet for design by this criterion is "bys" which enables 16% sequence recovery and 0.239 bits of mutual information with amino acid. See /projects/compbio/experiments/protein-predict/predict-2nd/compare-real/comparisons for more comparisons.
25 May	Memorial Day, no school
27 May	Energy functions for molecular modeling (such as AMBER and CHARMM): bond length, bond angle, torsion angle, clash, electrostatics, vague discussion of implicit solvent models
29 May	instructor evaluation forms; structure-strucutre alignment algorithms: iterated superposition and DALI distance maps.
1 Jun	Journal club: Michael Cusack Frank Alber, Friedrich Förster, Dmitry Korkin, Maya Topf, and Andrej Sali. Integrating Diverse Data for Structure Determination of Macromolecular Assemblies. Annual Review of Biochemistry 77:443-477 (July 2008) doi:10.1146/annurev.biochem.77.060407.135530 Journal Club: Kwok Chan Aleksandar Poleksic and Mark Fienup. Optimizing the size of the sequence profiles to increase the accuracy of protein sequence alignments generated by profile-profile algorithms. Bioinformatics 2008 24: 1145-1153. doi:10.1093/bioinformatics/btn097 http://bioinformatics.oxfordjournals.org/cgi/content/abstract/24/9/1145 ---------------------------------------------
3 Jun	Journal club: Rachel Bingham O. Alvizo, B. D. Allen and S. L. Mayo. Computational Protein Design Promises to Revolutionize Protein Engineering. Biotechniques, 42, 31-39, 2007. www.mayo.caltech.edu/publications/Alvizo.2007.Biotech.CPDrevolProtEng.pdf Journal Club: Jeff Ferguson Local Structural Plasticity of the Prion Protein. Analysis of NMR Relaxation Dynamics John H. Viles, David Donne, Gerard Kroon, Stanley B. Prusiner, Fred E. Cohen, H. Jane Dyson, and Peter E. Wright. Biochemistry 2001, 40, 2743-2753. doi:10.1021/bi002898a Get poster to BELS ---------------------------------------------
5 Jun	Final draft due Protein design: designing for structure or designing for function. Design for structure: goal is to check understanding of sequence/structure relationship. Initial designs were simple (coiled coil, 4-helix bundle). Most ambitious so far: top7. Explained how top7 scaffold was chosen and the RosettaDesign approach for getting the sequence and tweaking the initial backbone sketch. Design for function: use existing scaffold, tweak to change substrate or introduce new function. Discussed Hellinga group's use of maltose-binding protein, success at making biodetectors (such as for TNT), as well as retraction on the TIM-enzyme design. Discussed Baker group's enzyme design from scratch (using quantum chemistry to guess transition state for novel reaction). Also discussed use of directed evolution (phage display, ribosome display, noisy replication, cassette mutatgenesis) to do non-computational tweaking.
TUESDAY 9 Jun 4–7 p.m.	Poster Session

SoE home

Kevin Karplus's home page

Biomolecular Engineering Department

BME 220 home page

Karplus's lab page

UCSC Bioinformatics research

Questions about page content should be directed to

Kevin Karplus
Biomolecular Engineering
University of California, Santa Cruz
Santa Cruz, CA 95064
USA
karplus@soe.ucsc.edu
1-831-459-4250
318 Physical Sciences Building