Fri Jul 11 09:14:19 PDT 2008 T0498 Make started Fri Jul 11 09:14:52 PDT 2008 Running on cheep.cse.ucsc.edu Fri Jul 11 11:12:23 PDT 2008 Kevin Karplus I'm not sure what the point of this target is, since it is 59% identical to 2i38A, 2i2yA, ... and is just an NMR model, so not a super-high-resolution test. Of course, most of the NMR files aren't in the template library, but 2igdA (at 57% identity) is. Should I add some of the other hits and make more models? The top 4 or 5 models already are nearly indistinguishable. For try2, I'll just up clashes and breaks a lot, and see if we can make rosetta like the models also. Fri Jul 11 11:50:52 PDT 2008 Kevin Karplus For try3, I'm running again from alignments with lots of the non-template-library templates included, to see if it makes any difference. Fri Jul 11 11:55:23 PDT 2008 Kevin Karplus In try2, things haven't moved much, and L9 still clashes with V39. try2-opt3.gromacs0.repack-nonPC fixes that clash, but rosetta has kicked the tryptophan out of the core---a common problem with rosetta repacking. For the final polishing run, maybe I should start from gromacs optimized models, and increase the dry5 and dry6.5 weights a bit. Fri Jul 11 14:42:59 PDT 2008 Kevin Karplus try4-opt3 looks ok, and rosetta likes try4-opt3.gromacs0.repack-nonPC without having to flip W43 out of the core. The protein model may be a tiny bit too loose, but I can't be bothered to try to pack it any tighter. This is done except for metaserver results. Tue Jul 15 11:21:47 PDT 2008 Kevin Karplus MQAU likes pipe_int_TS1, SAM severs, Frankenstein_TS1, pro-sp3-TASSER_TS1, ... all with essentially the same predicted GDT (80.4%). MQAC likes MULTICOM-RAMK_TS2, Pcons_multi_TS5, MULTICOM-REFINE_TS3, SAM-T08-server_TS1, .... all with essentially the same predicted GDT (88.4%). I've started MQAU1 and MQAC1 metaserver runs. I think I'll also start an MQAX4 run, starting from all servers *except* SAM servers. Tue Jul 15 11:32:08 PDT 2008 Kevin Karplus Both MQAU1 and MQAC1 are favoring the SAM-T08-server_TS1 model, as expected. The MQAX4 run is favoring YASARA_TS3, which has no clashes or breaks. Tue Jul 15 13:04:38 PDT 2008 Kevin Karplus costfcn favorite models try1 MQAU1-opt3 MQAC1-opt3 try4-opt1 try2-opt3 MQAX4-opt3 try1-opt3 try3-opt3 try2 MQAX4-opt3 try4-opt3 MQAC1-opt3 MQAU1-opt3 try2-opt3 try3-opt3 try1-opt3 try3 MQAX4-opt3 try4-opt3 try3-opt3 MQAC1-opt3 try2-opt3 MQAU1-opt3 try1-opt3.gromacs0 try4 MQAX4-opt3 try4-opt3 MQAC1-opt3.gromacs0 try3-opt3.gromacs0 try2-opt3.gromacs0 try1-opt3.gromacs0 MQAU1-opt3.gromacs0.repack-nonPC rosetta MQAX4 try4 try2 try3 MQAC1 try1 MQAU1 (all -try3.gromacs0.repack-nonPC) Tue Jul 15 13:25:57 PDT 2008 Kevin Karplus Other than the position of F52 and some surface sidechains, all my models are nearly identical. Tue Jul 15 13:38:10 PDT 2008 Kevin Karplus Submitted with comment All my predictions, whether from SAM+undertaker or metaserver predictions, came out almost identical. There are some differences in placing F52 and some of the surface residues, but it is hard do distinguish among the models. The first two models are metaserver models, optimized by undertaker from YASARA_TS3. The first one is then energy-minimized by gromacs, and has the sidechains repacked by rosetta design. The third and fifth models is are optimized strictly from SAM+undertaker models. The 4th model is a meta-server model, but using our own SAM-T08 server. Model 1 T0498.MQAX4-opt3.gromacs0.repack-nonPC.pdb #< YASARA_TS3 # best rosetta energy 2 T0498.MQAX4-opt3.pdb #< YASARA_TS3 3 T0498.try4-opt3.pdb #< try3-opt3.gromacs0 < align(2gi9A) 4 T0498.MQAC1-opt3.gromacs0.pdb # < SAM-T08-server_TS1 5 T0498.try2-opt3.pdb # < try1-opt3 < align(2gi9A) Tue Jul 15 14:20:47 PDT 2008 Kevin Karplus After noticing some problems packing the aromatics in T0499, I looked at the best-models.pdb file here, but did not see similar problems. Thu Jul 17 17:01:00 PDT 2008 SAM-T08-MQAO hand QA T0498 Submitted Thu Jul 17 17:01:00 PDT 2008 SAM-T08-MQAU hand QA T0498 Submitted Thu Jul 17 17:01:01 PDT 2008 SAM-T08-MQAC hand QA T0498 Submitted