PFRMAT AL 
TARGET T0074 
AUTHOR 5827-4749-3439 
REMARK HMM prediction for t0074 
METHOD Note: this alignment was generated automatically, and has not 
METHOD been hand-edited, other than to truncate the alignment after 
METHOD residue 189 of the target (where the alignment begins to introduce 
METHOD very large gaps): 
METHOD 
METHOD T0074                     PWAVKPEDKAKYDAIFDSL-SPVNGFLSGDKVKPVL- 
METHOD 1OSA                      AEQLTEEQIAEFKEAFALFDKDGDGTITTKELGTVMR 
METHOD 
METHOD T0074                     -LNSKLPVDILGRVWELSDIDHDGMLDRDEFAVAM-- 
METHOD 1OSA                      SLGQNPTEAELQDMINEVDADGNGTIDFPEFLSLMAR 
METHOD 
METHOD T0074                     --------------FLVY--------CALEKEPVPMS 
METHOD 1OSA                      KMKEQDSEEELIEAFKVFDRDGNGLISAAELRHVMTN 
METHOD 
METHOD T0074                     LPPAL----VPPSKRK------------------TW 
METHOD 1OSA                      LGEKLTDDEVDEMIREADIDGDGHINYEEFVRMMVS 
METHOD 
METHOD As usual, we employed a two-pronged approach: (1) We scored the 
METHOD target against all the HMMs in our HMM library, and (2) we 
METHOD gathered homologs to the target using PsiBLAST, constructed an 
METHOD HMM for the target and homologs (using UCSC's HMM software), and 
METHOD scored PDB. This yielded two sets of scores, which we then used 
METHOD to find a target-structure match. 
METHOD 
METHOD The highest-scoring HMMs in the first stage of the analysis 
METHOD were a set of calcium-binding domains identified as belonging to the 
METHOD same superfamily in SCOP: d1tcob, d2sas, d2scpb, d2mysb,d1osa, and 
METHOD d1rro. The same proteins were given the highest scores by the HMM 
METHOD constructed for the target and homologs. Of these proteins, d1osa 
METHOD appeared to be the best match to the target. 
METHOD 
METHOD The alignment submitted was created by aligning the target and structure 
METHOD to an HMM constructed as follows: (1) converting a multiple sequence 
METHOD alignment  of 1osa and FASTA-identified homologs into an HMM using 
METHOD modelfromalign from the UCSC SAM software suite, (2) searching for 
METHOD homologous structures (and sequence homologs) using this HMM, 
METHOD (3) re-estimating the HMM using the original training sequences and 
METHOD new sequences found in step 2. 
MODEL  1 
PARENT 1osa 
P 121    A   1 
W 122    E   2 
A 123    Q   3 
V 124    L   4 
K 125    T   5 
P 126    E   6 
E 127    E   7 
D 128    Q   8 
K 129    I   9 
A 130    A  10 
K 131    E  11 
Y 132    F  12 
D 133    K  13 
A 134    E  14 
I 135    A  15 
F 136    F  16 
D 137    A  17 
S 138    L  18 
L 139    F  19 
S 140    K  21 
P 141    D  22 
V 142    G  23 
N 143    D  24 
G 144    G  25 
F 145    T  26 
L 146    I  27 
S 147    T  28 
G 148    T  29 
D 149    K  30 
K 150    E  31 
V 151    L  32 
K 152    G  33 
P 153    T  34 
V 154    V  35 
L 155    M  36 
L 156    L  39 
N 157    G  40 
S 158    Q  41 
K 159    N  42 
L 160    P  43 
P 161    T  44 
V 162    E  45 
D 163    A  46 
I 164    E  47 
L 165    L  48 
G 166    Q  49 
R 167    D  50 
V 168    M  51 
W 169    I  52 
E 170    N  53 
L 171    E  54 
S 172    V  55 
D 173    D  56 
I 174    A  57 
D 175    D  58 
H 176    G  59 
D 177    N  60 
G 178    G  61 
M 179    T  62 
L 180    I  63 
D 181    D  64 
R 182    F  65 
D 183    P  66 
E 184    E  67 
F 185    F  68 
A 186    L  69 
V 187    S  70 
A 188    L  71 
M 189    M  72 
TER 
END 
