Fri Aug 16 16:05:13 PDT 2002
t0195

Homology modeling 1jjfA c.69.1.2 or 1dqyA c.69.1.3

Tue Aug 20 13:59:14 PDT 2002

Best score on first run (after 21 hours) is try1-opt-scwrl.
Overall, it looks pretty good, but there is a bad break before K131
and after D153,  and it is still a bit foamy.

Turning up dry5, dry6.5, dry8, and breaks leaves try1-opt-scwrl at the top.

1 Sept 2002 Kevin Karplus	16:10

Best score now try2-opt.  There is a bad break between T158 and G159
which could probably be fixed by sliding part of the strand up two to pair
D153 with K181.

151>	fldnVAITgh
179>	ryKSCSAfa

Let's do another run starting from just the alignments with CB
constraints to encourage this pairing.

Tue Sep  3 10:05:36 PDT 2002	Kevin Karplus

try3-opt new best

V266-I268 not attached to sheet, somewhat awkward break at S182-C183.

Need to align Y265-K270 in parallel with R232-D239.

4 Sept 2002 Kevin Karplus	22:00

Made "conserve" script for conserved residues---no obvious active
site---conservation seems largely dictated by structural concerns.

Let's try sticking the strand back on with the following alignment

230>    ddrILIHVgdsd
263>      qdyVEIKkvhgfd


Thu Sep  5 17:24:35 PDT 2002	Kevin Karplus

In try4-opt, last strand has curled into helix.
Still have awkward break at C183-S184.  Need to rethink
constraints---I may have alignment wrong.

Final strand constraints need to extend to S40.


8 sept 2002 Kevin Karplus  

try5-opt still has bad break at C184-S184, but has the final strand
back in place.

Currently am trying to get alignment
155>	VAITGH
181>	KSCSAF

I should probably extend this back 2 to untwist the strand at the end.
I should pin this on the other side also:
153>	DNVAITG
179>	RYKSCSA
239<	DGVHILI


9 sept 2002 Kevin Karplus

I must have been VERY tired last night---I have the last strand
reversed there, which meant try6 was a total mess.
Trying again with try7 

// 153>    DNVAITG
// 179>    RYKSCSA
// 233>    ILIHVGD

and adding Hbonds as well as CB constraints, to provide more
"leverage" for the constraints to force the strand into position.

14:35

try7-opt does not have the big gap, but Hbonds are missing between the
strand containing F53 and the strand containing A156.  V155 and R179 clash

17:35

try8-opt has a nasty break between K181 and S182, but otherwise looks ok.
I should probably try to straighten out V3-S8, and attach them
anti-parallel to R13-N20.

I edited the try9 constraints to match the alignment I saw in try8,
and trimmed back the connections that seem to be pulling Y180 and K181
out of place.

Let's try reoptimizing with these constraints.

20:20 

Best scores with try9.constraints:	try9-opt, try8-opt, try5-opt,
					try6-opt, try7-opt, try4-opt

Best scores with no constraints:	try5-opt, try6-opt, try4-opt,
					try2-opt, try3-opt

Best scores with try5.constraints:	try5-opt, try6-opt, try8-opt,
					try9-opt, try4-opt, try7-opt

Best scores with try3.constraints:	try5-opt, try4-opt, try3-opt, try6-opt
					try9-opt, try8-opt, try2.1.100

Almost all the top models look very similar to me.

Let's submit

model 1:	try5-opt
model 2:	try9-opt
model 3:	try8-opt