Thu Jun 12 09:15:32 PDT 2008 T0449 Make started Thu Jun 12 09:15:48 PDT 2008 Running on cheep.cse.ucsc.edu Thu Jun 12 10:15:30 PDT 2008 Kevin Karplus The t06 multiple alignment contains 15 PDB sequences: 1snzA 1l7jA 1nsmA 1ns8A 1ns2A 1ns0A 1nsxA 1nsuA 1lurA 1z45A 2htaA 2cirA 2ciqA 1ygaA 1jovA The galactose mutarotase labeling of the target and several of the pdb sequences indicates that these are believed to be functionally similar proteins, thus true structures. They are essentially full-length matches (other than the C-terminal HIS tag). Thu Jun 12 15:40:48 PDT 2008 Kevin Karplus The b.30.5.4 domains do indeed score best, though some other families in the same superfamily also score fairly well (b.30.5.7, b.30.5.3, b.30.5.5). Thu Jun 12 21:05:55 PDT 2008 Kevin Karplus The try1-opt3 model looks pretty good and the top 3 aligments are all in agreement about phasing of the strands and helices. The loop from N182-I193 seems to be different in the top 3 alignments than in try1, and the first alignment has a better loop from I206 to N214. Perhaps I should run try2 from alignments, but with sheet and helix constraints from try1. I should also concentrate on the top 7 templates: 1nszA 1snzA 1lurA 2cirA 1ygaA 2htaA 1z45A C251, E253, H143, H73 are conserved and look like they may form part of the active site. I may want to constrain them based on top alignments, but I won't do that for try2. There is probably a salt-bridge for H143.ND1 to E253.OE2, for example. Thu Jun 19 12:00:23 PDT 2008 SAM-T08-MQAO hand QA T0449 Submitted Thu Jun 19 12:00:23 PDT 2008 SAM-T08-MQAU hand QA T0449 Submitted Thu Jun 19 12:00:23 PDT 2008 SAM-T08-MQAC hand QA T0449 Submitted Sun Jun 29 20:40:02 PDT 2008 Kevin Karplus The try1 costfcn likes MQAC1-opt3.gromacs0.repack-nonPC MQAU1-opt3.gromacs0 try2-opt3.repack-nonPC try1-opt3.repack-nonPC The try2 costfcn likes MQAC1-opt3.gromacs0.repack-nonPC MQAU1-opt3.gromacs0 try2-opt3.gromacs0.repack-nonPC try1-opt3.gromacs0.repack-nonPC It looks like it is time for a polishing run to reduce clashes and breaks. There is a difference in phase for strand L121-D133 between MQAC1 and the rest. I think I like the alignment the rest get somewhat better, but I'll have to see after try3. try3 will try improving the MQAC1 alignment try4 will try improving the other alignments. Mon Jun 30 09:51:44 PDT 2008 Kevin Karplus Even with the try4 costfcn, try3-opt3 scores best. try4 still has bad breaks and clashes. Naturally rosetta likes the low-clash model try3-opt3.gromacs0.repack-nonPC best. Tue Jul 1 13:39:21 PDT 2008 Kevin Karplus try5-opt3 improves on try4-opt3, but is still beat by try3-opt3, probably becasue of the breaks in try5. These look like several adjacent breaks, of the sort introduced by gromacs. They're pretty big, so I need to do a polishing run with breaks turned up a lot. Tue Jul 1 13:51:06 PDT 2008 Kevin Karplus try6 to optimize try5 started on peep. Tue Jul 1 18:02:52 PDT 2008 Kevin Karplus I don't seem to be able to make any progress on closing the many bad breaks in try6. None of the gap-closing operators is ever succeeding. Maybe I should just do a polishing run on the try3 model and call it quits. Tue Jul 1 18:10:56 PDT 2008 Kevin Karplus Started try7 to polish try3. Maybe I'll make one more attempt to close gaps in try6. (Started try8 to aggressively close gaps in try6). Tue Jul 1 22:05:04 PDT 2008 Kevin Karplus try7 ran ok, but try8 crashed with undertaker: Spot.cc:172: XYZpoint Spot::get_spot(const ChainsResiduesAndAtoms*, int, const Pointlist&, const Conformation*, bool, const Spot*) const: Assertion `return_point.mag2() < 1.e12' failed. Wed Jul 2 09:23:18 PDT 2008 Kevin Karplus try9, which was similar to try8, ran ok and went a long way toward closing gaps. try10 which polished try9-opt3.gromacs0, reduced breaks (but not clashes) from try9-opt3.gromacs0, which had already reduced clashes considerably from try9-opt3. I like the strand N120-D133 better in try10, but try7 has some other parts that are better. It might be worth making a chimera, mostly from try10, but F81-L95 from try7 Wed Jul 2 09:55:18 PDT 2008 Kevin Karplus try11 started on peep to optimize chimera-try10-try7 Closing gaps on this target was very slow. I wonder what the problem was? too high a clash cost to allow things to move? Erroneous sheet constraints that are holding things in difficult positions?? Wed Jul 2 12:43:34 PDT 2008 Kevin Karplus try11 scores well, and I'm running out of time, so I'll make it the first model, though it is quite close to the try10 model. ReadConformPDB T0449.try11-opt3.pdb # < chimera-try10-try7 chimera-try10-try7: mostly from try10-opt3 F81-L95 from try7-opt3 ReadConformPDB T0449.try7-opt3.pdb # < try3-opt3 < MQAC1-opt3.gromacs0.repack-nonPC < MULTICOM-REFINE_TS5 ReadConformPDB T0449.try10-opt3.pdb #