PFRMAT AL 
TARGET T0076 
AUTHOR 9070-5088-8627 
METHOD Overview 
METHOD  
METHOD Fold recognition was performed using the Target98 (SAM-T98) method 
METHOD [3] using SAM version 2.1.1 [1], a refinement of the methods developed 
METHOD by this group for CASP2 [2].  This method attempts to find and multiply  
METHOD align a set of homologs to a given sequence, then create an HMM from that  
METHOD multiple alignment. 
METHOD  
METHOD First, a set of sequence weights is determined from the alignment.  Next,  
METHOD Modelfromalign is used to build the model from the alignment and the  
METHOD sequence weights.  Finally, hmmscore performs a local, all-paths scoring  
METHOD of the sequences, using a reversed-sequence normalization feature. 
METHOD  
METHOD The weighting method, detailed in upcoming publications [3,4], 
METHOD combines the Henikoffs' scheme [5], Dirichlet mixtures [6], and an 
METHOD entropy method to set the final weights. 
METHOD  
METHOD Alignment generation 
METHOD  
METHOD The initial step uses BLASTP to search NRP twice: once to produce a set 
METHOD of very close homologs, and once to produce a set of possible homologs. 
METHOD  
METHOD The method then uses multiple iterations of a selection, training, and  
METHOD alignment procedure.  Each iteration involves an initial alignment, a set  
METHOD of search sequences, a threshold value, and a transition regularizer.  
METHOD  
METHOD The first iteration uses a single sequence (or seed alignment) as the  
METHOD initial alignment and the close homologs found by BLASTP are used as the  
METHOD search set.  The threshold is set very strictly, so that only good matches  
METHOD to the sequence are considered.  This iteration uses a transition regularizer  
METHOD that was designed to match the gap costs used by BLASTP. 
METHOD  
METHOD On subsequent iterations the input alignment is the output from the 
METHOD previous iteration, the search set is the larger set of possible 
METHOD homologs found by BLASTP, and the thresholds are gradually loosened. 
METHOD The second through second-from-last iteration use a ``long-match'' 
METHOD transition regularizer, and the final iteration uses a transition regularizer  
METHOD trained on FSSP alignments. 
METHOD  
METHOD References 
METHOD [1] R. Hughey and A. Krogh, CABIOS 12(2): 95-107, 1996. 
METHOD     http://www.cse.ucsc.edu/research/compbio/sam.html.   
METHOD [2] K. Karplus, K. Sjolander, C. Barrett, M. Cline, D. Haussler, R. 
METHOD     Hughey, L. Holm, and C. Sander, Proteins: Structure, Function, and  
METHOD     Genetics, Suppl. 1, 134-9, 1997. 
METHOD [3] K. Karplus, C. Barrett, and R. Hughey, Technical Report UCSC-CRL-98-06, 
METHOD     Department of Computer Engineering, Univ. of California, Santa Cruz, 1998. 
METHOD [4] J. Park, K. Karplus, C. Barrett, R. Hughey, D. Haussler, T. Hubbard, 
METHOD     and C. Chothia, http://cyrah.med.harvard.edu/~jong/assess_final.html, 1998. 
METHOD [5] S. Henikoff and J. C. Henikoff, JMB, vol 243, pp 574-578, Nov 1994. 
METHOD [6] K. Sjolander, K. Karplus, M. P. Brown, R. Hughey, A. Krogh, I. S. 
METHOD    Mian, and D. Haussler, CABIOS 12(4):327-345, 1996. 
METHOD  
METHOD  
METHOD T0076 clearly has 2 EF-hand domains, with very high similarity to 
METHOD calmodulins (such as 1deg). 
METHOD  
METHOD Because the problem is to distinguish which of the many EF-hand 
METHOD domains is the closest model, we chose to use only the second 
METHOD iteration of the target-alignment-building process for our target HMM. 
METHOD  
METHOD The calmodulins provide a good fit to the EF-hand domains, but the 
METHOD central helix would be disrupted by the prolines and glycines in 
METHOD PNGFDMPGDP.   Myosins are a much better fit for the central helix, 
METHOD particularly 2mysC.  Unfortunately, 2mysC has a chain-tracing error in 
METHOD the first domain, so is a bit dangerous to use as a basis for homology 
METHOD modeling. 
METHOD  
METHOD We finally decided to use 1almC, a theoretical model based on 2mysC, 
METHOD which seems to fix the chain-tracing error. We unaligned the second 
METHOD binding-pocket of the first EF-hand, since it is much shorter in 
METHOD T0076, and 1almC seems to be untrustworthy in that loop.  None of the 
METHOD other calmodulin-like domains we looked at had a loop that we could 
METHOD confidently plug in to fill the gap, so we just left it unaligned. 
MODEL 1 
PARENT 1alm_C 
T 2 E 5 
D 3 Q 6 
D 4 Q 7 
S 5 D 8 
P 6 D 9 
Y 7 F 10 
K 8 K 11 
Q 9 E 12 
A 10 A 13 
F 11 F 14 
S 12 L 15 
L 13 L 16 
F 14 F 17 
D 15 D 18 
R 16 R 19 
H 17 T 20 
G 18 G 21 
T 19 D 22 
G 20 A 23 
R 21 K 24 
I 22 I 25 
P 23 T 26 
K 24 L 27 
T 25 S 28 
S 26 Q 29 
I 27 V 30 
G 28 G 31 
D 29 D 32 
L 30 I 33 
L 31 V 34 
R 32 R 35 
A 33 A 36 
C 34 L 37 
G 35 G 38 
Q 36 Q 39 
N 37 N 40 
P 38 P 41 
T 39 T 42 
L 40 N 43 
A 41 A 44 
E 42 E 45 
I 43 I 46 
T 44 N 47 
E 45 K 48 
I 46 I 49 
E 47 L 50 
V 54 I 63 
D 55 T 64 
M 56 F 65 
E 57 E 66 
Q 58 E 67 
F 59 F 68 
L 60 L 69 
Q 61 P 70 
V 62 M 71 
L 63 L 72 
N 64 Q 73 
R 65 A 74 
P 66 A 75 
N 67 A 76 
G 68 N 77 
F 69 N 78 
D 70 K 79 
M 71 D 80 
P 72 Q 81 
G 73 G 82 
D 74 T 83 
P 75 F 84 
E 76 E 85 
E 77 D 86 
F 78 F 87 
V 79 V 88 
K 80 E 89 
G 81 G 90 
F 82 L 91 
Q 83 R 92 
V 84 V 93 
F 85 F 94 
D 86 D 95 
K 87 K 96 
D 88 E 97 
A 89 G 98 
T 90 N 99 
G 91 G 100 
M 92 T 101 
I 93 V 102 
G 94 M 103 
V 95 G 104 
G 96 A 105 
E 97 E 106 
L 98 L 107 
R 99 R 108 
Y 100 H 109 
V 101 V 110 
L 102 L 111 
T 103 A 112 
S 104 T 113 
L 105 L 114 
G 106 G 115 
E 107 E 116 
K 108 K 117 
L 109 M 118 
S 110 T 119 
N 111 E 120 
E 112 E 121 
E 113 E 122 
M 114 V 123 
D 115 E 124 
E 116 E 125 
L 117 L 126 
L 118 M 127 
K 119 K 128 
G 120 G 129 
V 121 Q 130 
P 122 E 131 
V 123 D 132 
K 124 S 133 
D 125 N 134 
G 126 G 135 
M 127 C 136 
V 128 I 137 
N 129 N 138 
Y 130 Y 139 
H 131 E 140 
D 132 A 141 
F 133 F 142 
V 134 V 143 
Q 135 K 144 
M 136 H 145 
I 137 I 146 
L 138 M 147 
A 139 S 148 
N 140 V 149 
TER 
END 
