Kevin Karplus	24 May 1998

Target t57 is clearly 1gaeO, which has fssp rep 1gypA, though 1dapA is
a better score in the target98 models than 1gypA.  The 1gypA homologs
score much better with the target model, perhaps because some of them are
included in the t57.t98_6 alignment.

The 1dapA-t57-global alignment picks up an excellent 39-residue alignment
at the beginning and some so-so alignments later on.  The fssp
alignment doesn't score as well, but is worth looking at to see if
structure is better preserved.

Kevin Karplus 26 May 1998

The 1gypA-t57-fssp-global alignment looks a bit better than
1dapA-t57-fssp-global, though there are still some very suspect regions.

Kevin Karplus 16 July 1998

Remade the alignments using t57.t98_5 as the base alignment for
t57-... alignments. The t57-1gadP-post alignment looks good, but
leaves about 80 residues of t57 unaligned.  I'll have to check
compatibility with the alignments to the fssp rep 1gypA.

28 July 1998

wu-blast likes 1dapA best -4.07
double-blast likes

best-score non-self alignment:

1gadP/t57-1gadP-global	1gadP             330       -41.35      -44.07
1gadP/t57-1gadP-post	1gadP             330       -42.73      -44.07
1dapA/1dapA-t57-global	T0057             340       -24.63      -29.52
1dapA/1dapA-t57-post	T0057             340       -26.02      -29.52
1gypA/1gypA-t57-const-global	T0057             340       -27.03      -28.36
1dapA/1dapA-t57-vit	T0057             340       -31.82      -26.85
1gypA/1gypA-t57-global	T0057             340       -23.13      -26.07
1gypA/1gypA-t57-post	T0057             340       -24.52      -26.07
1gadP/t57-1gadP-vit	1gadP             330       -30.93      -25.61
1gypA/1gypA-t57-fssp-global	T0057             340       -22.91      -24.52
1dapA/t57-1dapA-global	1dapA             320       -18.41      -23.03
1dapA/t57-1dapA-post	1dapA             320       -19.80      -23.03
1dapA/t57-1dapA-vit	1dapA             320       -25.02      -20.57
1dapA/1dapA-t57-const-global	T0057             340       -22.16      -20.44
1gypA/t57-1gypA-global	1gypA             358       -14.13      -17.19

I made a first attempt at a hand alignment in t57-1gadP-hand.a2m,
mainly sweeping closed the big gaps, and unaligning one unbridgeable
section where a helix seems to be missing.


Mon Aug  3 10:55:03 PDT 1998

I looked at the 1gadP alignments and fussed around with them a bit,
coming up with t57-1gadp-hand2.a2m and t57-1gadp-hand3.a2m.


Here is a draft of the results:

Wu-blast found 1dapA as a reasonably close homolog, but double-blast
found 1gadP (with intermediate sequence G3P_METBR from Swissprot) to
be a better match.  Our target HMM scored 1gadP best, but the library
HMM scored 1dapA best.  The 1gadP sequence does not have a model in
our library, though 1gypA (the FSSP representative for 1gadP) scored
fairly well.  Summing both ways, the top hits were 1gd1O, 1dapA, and
1gadP, all with scores well into the region where we had fewer than 1%
false positives in our superfamily-recognition tests.

The 1gd1O and 1gadP alignments looked OK, but needed extensive hand
editing.  We finally chose a 1gd1O alignment as having the most
reasonable gaps.  The binding site for NAD is well preserved in both,
as is the beta sheet for the dimerization site.