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. New this year, we are also occasionally using ProteinShop to manipulate proteins by hand, to produce starting points for undertaker optimization. We expect this to be most useful in new-fold all-alpha proteins, where undertaker often gets trapped in poor local minima by extending helices too far. Another 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. T0374 got moderate hits to 2f14A and 2fe7A, and the HMMs got lots of hits to the d.108.1.1 SCOP domain (top 2ge3A). All of the alignments were in agreement except for some disagreement over the C-terminal, where t2k was predicting a helix and all others predicting a strand in the secondary structure. In all models there is a bulge in the beta strand at residues D74-A84. We attempted to remove the bulge, but were unable to do so. Model 1 is T0374.try14-opt2, our highest scoring model using Undertaker's unconstrained.costfcn cost function. This model came from try12-opt2 < try11-opt2.gromacs0. try11-opt2.gromacs0 came from an Undertaker run deriving a model from alignments (top alignment 1tiqA), applying the sheet and helix constraints from try6-opt2, and then re-optimizing with gromacs0. We chose try6-opt2 sheet and helix constraints because it had formed a nicer beta-sheet than previous models. Model 2 is T0374.try15-opt2, our second highest scoring model using Undertaker's unconstrained.costfcn cost function. This model is a slight variation on try14-opt2, based also on the try12-opt2 model, but aimed at improving breaks, clashes, and hydrophobic packing. Model 3 is T0374.try13-opt2.gromacs0.repack-nonPC, is not a great scoring model according to Undertaker, but its Rosetta-optimized structure, T0374.try13-opt2.gromacs0.repack-nonPC, is the highest scoring model according to Rosetta. try13-opt2 came from an optimization from all models, coming up with the then best current model try2-opt2 to start its optimization, then re-optimized by gromacs and repacked (except for PRO and CYS) by Rosetta. try2-opt2 was made during an Undertaker run starting from alignments (top alignment 1tiqA) using the sheet and helix constraints of try1-opt2, which seemed like reasonable constraints. However, there was disagreement over the C-terminal strand where the t2k predicted a helix, but all others believed it to be a strand. Model 4 is T0374.try6-opt2, a not terribly great scoring model according to Undertaker, but was better liked by Rosetta after its repacking of the sidechains. try6-opt2 is included here since it was one of the first models that had the beginnings of a well-formed edge strand. try6-opt2 was made from an Undertaker optimization of try3-opt2 < try2-opt2. try2-opt2 was made during an Undertaker run starting from alignments (top alignment 1tiqA) using the try1-opt2 sheet and helix constraints. Model 5 is T0374.try3-opt2.gromacs0.repack-nonPC is included mainly for a little bit of variety. It has average scores according to both Undertaker's unconstrained.costfcn cost function and Rosetta's cost function. It has a C-terminal helix where most of our models have a strand. This helix is predicted by t2k but is predicted to be strand by the others. try3-opt2 was an optimization of try2-opt2, then re-optimized by gromacs and repacked (except for PRO and CYS) by rosetta. try2-opt2 was made by Undertaker from alignments (top alignment 1tiqA) using the sheet and helix constraints from the try1-opt2 automatic model.