The SAM-T06 hand predictions use methods similar to SAM_T04 in CASP6 and the SAM-T02 method in CASP5. We start with a fully automated method (implemented as the SAM_T06 server): Use the SAM-T2K and SAM-T04 methods for finding homologs of the target and aligning them. The hand method also uses the experimental new SAM-T06 alignment method, which we hope is both more sensitive and lass prone to contamination by unrelated sequences. Make local structure predictions using neural nets and the multiple alignments. We currently use 10 local-structure alphabets: DSSP STRIDE STR2 an extended version of DSSP that splits the beta strands into multiple classes (parallel/antiparallel/mixed, edge/center) ALPHA an discretization of the alpha torsion angle: CA(i-i), CA(i), CA(i+1), CA(i+2) BYS a discretization of Ramachandran plots, due to Bystroff CB_burial_14_7 a 7-state discretization of the number of C_beta atoms in a 14 Angstrom radius sphere around the C_beta. near-backbone-11 an 11-state discretization of the number of residues (represented by near-backbone points) in a 9.65 Angstrom radius sphere around the sidechain proxy spot for the residue. DSSP_EHL2 CASP's collapse of the DSSP alphabet DSSP_EHL2 is not predicted directly by a neural net, but is computed as a weighted average of the other backbone alphabet predictions. O_NOTOR2 an alphabet for predicting characteristics of hydrogen bonds from the carbonyl oxygen N_NOTOR2 an alphabet for predicting characteristics of hydrogen bonds from the amide nitrogen We hope to add more networks for other alphabets over the summer. We make 2-track HMMs with each alphabet (1.0 amino acid + 0.3 local structure) and use them to score a template library of about 8000 (t06), 10000 (t04), or 15000 (t2k) templates. The template libraries are expanded weekly, but old template HMMs are not rebuilt. We also used a single-track HMM to score not just the template library, but a non-redundant copy of the entire PDB. One-track HMMs built from the template library multiple alignments were used to score the target sequence. All the logs of e-values were combined in a weighted average (with rather arbitrary weights, since we still have not taken the time to optimize them), and the best templates ranked. Alignments of the target to the top templates were made using several different alignment methods (mainly using the SAM hmmscore program, but a few alignments were made with Bob Edgar's MUSCLE profile-profile aligner). Generate fragments (short 9-residue alignments for each position) using SAM's "fragfinder" program and the 3-track HMM which tested best for alignment. Residue-residue contact predictions are made using mutual information, pairwise contact potentials, joint entropy, and other signals combined by a neural net. The contact prediction method is expected to evolve over the summer, as new features are selected and new networks trained. Then the "undertaker" program (named because it optimizes burial) is used to try to combine the alignments and the fragments into a consistent 3D model. No single alignment or parent template was used as a frozen core, though in many cases one had much more influence than the others. The alignment scores were not passed to undertaker, but were used only to pick the set of alignments and fragments that undertaker would see. Helix and strand constraints generated from the secondary-structure predictions are passed to undertaker to use in the cost function, as are the residue-residue contact prediction. One important change in this server over previous methods is that sheet constraints are extracted from the top few alignments and passed to undertaker. After the automatic prediction is done, we examine it by hand and try to fix any flaws that we see. This generally involves rerunning undertaker with new cost functions, increasing the weights for features we want to see and decreasing the weights where we think the optimization has gone overboard. Sometimes we will add new templates or remove ones that we think are misleading the optimization process. One new trick is to optimize models with gromacs to knock them out of a local minimum. The gromacs optimization does terrible things to the model (messing up sidechains and peptide planes), but is good at removing clashes. The resulting models are only a small distance from the pre-optimization models, but score much worse with the undertaker cost functions, so undertaker can move them more freely than models it has optimized itself. We had excellent hits to 1zjjA (not too surprising as even BLAST found it with E-value 1.2e-17). Closing the loops turned out to be harder than expected, but turning up the break costs allowed undertaker to close most of the gaps. There was some disagreement between different models about how to handle the helix F146-R159. Model 1 is try15-opt2, the best-scoring model with the undertaker try15.costfcn cost function. It was optimized by undertaker from a chimera of try14-opt2 and try12-opt2, with F143-I164 taken from try12-opt2, and the rest from try14-opt2. Model 2 is try15-opt2.gromacs0.repack-nonPC, the best-scoring model with rosetta's cost function, of those backbones for which rosetta repacked sidechains. It was created from try15-opt2 by reoptimizing with gromacs then repacking sidechains (except PRO and CYS) by rosetta. Model 3 is try14-opt2, the base for the chimera optimized to make try15-opt2. It was optimized by undertaker from try13-opt2.gromacs0. try13-opt2 < try8-opt2 < try6-opt2 < SAM_T06_server_TS1 (Note: the SAM_T06 server is our server, and uses a slight variant of the method use for our hand SAM-T06 predictions, so we feel no guilt over using it as a starting point.) Model 4 is try12-opt2, which messed up the termini and has bad gaps, but scores best with undertaker's default "unconstrained" cost function (which does not have sufficient penalty for breaks for doing such close homology modeling. try12-opt2 < try11-opt2 < try10-opt2 < alignments (mainly 1zjjA) Model 5 is sidechain replacement by SCWRL on an alignment to 1zjjA.