Tue Jun 29 09:01:31 PDT 2004
T0221

Due 10 Aug

Tue Jun 29 12:55:35 PDT 2004	Kevin Karplus

T0221 looks like a comparative model with a.28.1.* as the templates.

Tue Jun 29 19:05:56 PDT 2004	Kevin Karplus

try1-opt2 looks pretty good.  The conserved hydrophobics are buried
and the conserved charged residues are exposed on the end.  The only
de novo part is residues 1-8, which MAY continue the initial helix.

We may just want to do some polishing on this one to try to get
slightly tighter packing.

Wed Jun 30 01:13:47 PDT 2004	Kevin Karplus

try2-opt2 is new best scorer (with the try1 costfcn).
try2-opt2.repack-nonPC looks a little better to Rosetta than
try1-opt2.repack-nonPC, though undertaker thinks it is the worst so far.

This model seems almost done.  I'll do one polishing round without
constraints and with dry weights turned up a bit higher for tighter packing.

Without constraints (try3.costfcn), the best scorer is still
try2-opt2, but the try2-opt2.repack-nonPC does not look quite as bad
as with the try1 costfcn.

Wed Jun 30 07:34:40 PDT 2004	Kevin Karplus

try3-opt2 is the new best scoring model, but it does not do much
better than try2-opt2, and rosetta preferes try2-opt2.repack-nonPC,
because of slightly better Erep (fewer clashes) that is not
quite compensated for by better terms elsewhere.

The residues Rosetta does not like after repacking are
res aa    Eatr     Erep    Esol  Eh2o_sol   Eaa    Edun  Eintra    Ehbnd   Epair   Eref   Eh2o   Eh2o_hb  Eres
other-bump: 2.92307 Ang (T0221)L42.CD2 and (T0221)A47.CA threshold= 3.17578 cost= 0.391955
other-bump: 2.99508 Ang (T0221)L42.CD2 and (T0221)A47.N threshold= 3.1578 cost= 0.271982
other-bump: 2.98017 Ang (T0221)L42.CD2 and (T0221)A47.CB threshold= 3.07655 cost= 0.173845
 42 LEU    -3.4    17.8     0.9     0.0    -0.1     2.4     0.5    -0.1     0.0    -0.1     0.0     0.0    18.1
 47 ALA    -3.4    16.6     1.5     0.0    -0.3    -0.0     0.0    -0.7     0.0     0.1     0.0     0.0    13.6

other-bump: 2.87966 Ang (T0221)F20.CZ and (T0221)F37.CD1 threshold= 3.21103 cost= 0.47979
other-bump: 2.89195 Ang (T0221)F20.CZ and (T0221)L24.CD2 threshold= 3.17246 cost= 0.426192
 37 PHE    -5.4     3.8     1.5     0.0     0.0     2.2     4.7    -1.1     0.0    -1.0     0.0     0.0     6.8
 24 LEU    -4.2     6.1     0.9     0.0    -0.2     2.0     0.1    -0.8     0.0    -0.1     0.0     0.0     4.0

Rosetta does not see what undertaker thinks is the worst clash:
neighbor-bump: 3.12636 Ang (T0221)W57.CD1 and (T0221)L58.CD2 threshold= 3.54883 cost= 0.532567
but this may be because undertaker is overly strict on TRP,
particularly in neighbor positions, because of too-small samples when
setting the parameters.

Undertaker does see one worse clash in try3-opt2:
other-bump: 2.10408 Ang (T0221)L42.CD1 and (T0221)K46.CB threshold= 3.05865 cost= 0.894028
The change in rotamer for L42 spreads the errors out in the repacking,
which may be better, but looks worse to undertaker.

Other than adding comments, I think we're ready to submit.  I think
this is about as polished as the model gets.


Fri Jul 16 12:50:35 PDT 2004 Kevin Karplus

Maybe I'll do one more short polishing run, then submit.