Thu Jun  5 09:49:50 PDT 2008
T0434
Make started Thu Jun  5 09:50:24 PDT 2008
Running on cheep.cse.ucsc.edu

Thu Jun  5 09:52:25 PDT 2008 Kevin Karplus

This looks like a homology model, with pdb-blast getting an e-value of
9e-15 to 2jv0A and 1e-14 to 2qpwA (apparently identical alignments).
This covers residues 81-182, so there may be some additional
non-homology work to do, or it may be that pdb-blast just has
too-short alignments.

Thu Jun  5 12:07:53 PDT 2008 Kevin Karplus

The HMMs get a clear signal for b.85.7.* superfamily, particularly b.85.7.1.
2qpwA comes up particularly strong in the t2k search (which has
smaller radius).

Thu Jun  5 13:04:40 PDT 2008 Kevin Karplus

The t06 and t04 HMMs prefer 2rfiA, but have 2qpwA in 5th place on the list.
Since the top hits are not in SCOP, I looked them up in VAST, and they
all seem to be instances of the SET superfamily:
"SET domain. SET domains are protein lysine methyltransferase
enzymes. SET domains appear to be protein-protein interaction
domains. It has been demonstrated that SET domains mediate
interactions with a family of proteins that display similarity with
dual-specificity phosphatases (dsPTPases). A subset of SET domains
have been called PR domains. These domains are divergent in sequence
from other SET domains, but also appear to mediate protein-protein
interaction. The SET domain consists of two regions known as SET-N and
SET-C. SET-C forms an unusual and conserved knot-like structure of
probably functional importance. Additionally to SET-N and SET-C, an
insert region (SET-I) and flanking regions of high structural
variability form part of the overall structure."

Mon Jun  9 22:04:08 PDT 2008 Kevin Karplus

MQAC1-opt3 is based mainly on Pcons_multi_TS4
MQAU1-opt3 is based mainly on SAM-T08-server_TS1

With try1.costfcn, the best models are MQAC1-opt3 try1-opt3.gromacs0 MQAU1-opt3

These models are all fairly similar, with slight differences in the N
and C termini.

I'll do a polishing run from all models.

Sat Jun 14 15:22:10 PDT 2008 Kevin Karplus

try2-opt3 scores best with try2 costfcn and is based mainly on try1-opt3.gromacs0.repack-nonPC
Rosetta's favorite (though it is still not fond of it) is try2-opt3.gromacs0.repack-nonPC

Perhaps I should do another polishing run, starting from the gromacs
optimized models.

Sat Jun 14 21:09:29 PDT 2008 Kevin Karplus

try3-opt3 (optimized from try2-opt3.gromacs0) now scores best with the try3 costfcn.
Rosetta likes best try3-opt3.gromacs0.repack-nonPC.

I don't think I'll get much more on this homology model, so I might as
well submit
model
1	T0434.try3-opt3.pdb	< try2-opt3.gromacs0 < MQAU1-opt3 < SAM-T08-server_TS1
2	T0434.try2-opt3.pdb	< MQAU1-opt3 < SAM-T08-server_TS1
3	T0434.MQAC1-opt3.pdb	< Pcons_multi_TS4
4	T0434.try1-opt3.gromacs0.pdb < 2rfiA
5	T0434.MQAU1-opt3.pdb	< SAM-T08-server_TS1


Thu Jun 19 12:00:16 PDT 2008   SAM-T08-MQAO hand QA T0434 Submitted
Thu Jun 19 12:00:17 PDT 2008   SAM-T08-MQAU hand QA T0434 Submitted
Thu Jun 19 12:00:17 PDT 2008   SAM-T08-MQAC hand QA T0434 Submitted