Mon Jul 12 10:45:54 PDT 2004
T0236

DUE 13 Aug

Mon Jul 12 18:41:54 PDT 2004	Kevin Karplus
This gets only weak hits to alpha-helical proteins (best is 1r5qA),
but those hits seem to get most of the structure (two helices crossing
2 helices).

In try1, undertaker may be bending the C-terminal helix at I93
inappropriately.  Actually, that isn't such a bad place to turn (G92),
and there might be a turn at G81 also.

This may be a dimer, since there seems to be a single somewhat
hydrophobic face.

Interaction between conserved Y11 and conserved H51 looks good. 

The conserved Y80 is adjacent to the conserved P44 also, and the OH
sticks out the end.  I wonder if G81 IS supposed to be a turn, so that
the Y80.OH is more exposed.  The position of Y84 is not bad though, so
the helix may go straight through. 

Since undertaker usually has no trouble creating helices, for try2
I'll simply turn off the constraints and see what happens.


Fri Sep 24 21:31:22 PDT 2004 Kevin Karplus

Using smooth_GDT we get
name			length	missing_atoms	rmsd	rmsd_ca	GDT	smooth_GDT
robetta-model5.pdb.gz	110	 0.0000		 5.8902	 5.2701	-58.3333	-53.7096
T0236.try1-opt1.pdb.gz	110	 0.0000		 6.3721	 5.3907	-53.2738	-51.1934
T0236.try1-opt2.pdb.gz	110	 0.0000		 6.4028	 5.3969	-53.8690	-51.0781	full auto
robetta-model1.pdb.gz	110	 0.0000		 8.0652	 7.5569	-52.6786	-49.9510
robetta-model2.pdb.gz	110	 0.0000		 8.3086	 7.4033	-50.8929	-49.0355
robetta-model3.pdb.gz	110	 0.0000		 8.6113	 7.7456	-51.4881	-48.7155
robetta-model4.pdb.gz	110	 0.0000		 8.7978	 7.9237	-50.2976	-48.0923

This was a cancelled target, so we didn't submit anything.
Our fully automatic run was competitive with robetta.