PFRMAT AL 
TARGET T0046 
AUTHOR 5827-4749-3439 
METHOD T0046 structure prediction 
METHOD Note: this alignment was generated automatically, and we opted not 
METHOD to try to align the C-terminal strand. 
METHOD 
METHOD As usual, we employed a two-pronged approach: (1)  
METHOD  We scored the target against all the HMMs and threading models  
METHOD in our libraries,  
METHOD and (2) we gathered homologs to the target using PsiBLAST,  
METHOD constructed an HMM for the target and homologs (using UCSC's HMM  
METHOD software), and scored PDB.  
METHOD This yielded three sets of scores, which we then used  
METHOD to find a target-structure match. 
METHOD  
METHOD The structure 2mcm was the second highest scoring structure 
METHOD for the HMM constructed from the target and homologs, and 
METHOD also scored well among  our threading model library. 
METHOD The highest scoring threading model was (scop id) d1ggta3, 
METHOD an immunoglobulin-like fold.  This same model was ranked 
METHOD among the top 3 structures for all of the T46 sequence homologs. 
METHOD The structure of 1ggta3 is similar to 2mcm, and this focused 
METHOD our attention on 2mcm, among the highest-scoring but weak HMM 
METHOD hits.  We decided to use the HMM alignment since it showed greater 
METHOD sequence similarity. 
METHOD  
METHOD The alignment submitted was created by aligning the target and structure  
METHOD to the HMM constructed for the target and homologs as follows: (1)  
METHOD constructing an HMM for T0046 alone using modelfromalign 
METHOD and the target sequence as the template alignment,  
METHOD (2) reestimating the HMM using sequences found using iterated FASTA 
METHOD (using new sequences as seeds in the search until convergence) 
METHOD (3) then reestimating that HMM using sequence weighting, 
METHOD to allow the HMM to generalize. The weighting method 
METHOD used Henikoff weighting to set relative weights and 
METHOD our own method of estimating the number of independent 
METHOD observations to set the total weight allotted 
METHOD the sequences. This alignment submitted was not edited in any way 
METHOD from the automatic alignment. 
MODEL 1 
PARENT 2mcm 
I    1    A    1 
P    2    P    2 
S    3    G    3 
I    4    V    4 
T    5    T    5 
A    6    V    6 
Y    7    T    7 
S    8    P    8 
K    9    A    9 
N    10    T    10 
G    11    G    11 
L    12    L    12 
K    13    S    13 
I    14    N    14 
E    15    G    15 
F    16    Q    16 
T    17    T    17 
F    18    V    18 
E    19    T    19 
R    20    V    20 
S    21    S    21 
N    22    A    22 
T    23    T    23 
N    24    G    24 
P    25    L    25 
S    26    T    26 
L    40    P    27 
D    41    G    28 
M    42    T    29 
T    43    V    30 
D    44    Y    31 
F    45    H    32 
V    46    V    33 
F    47    G    34 
Q    48    Q    35 
A    49    C    36 
A    50    A    37 
V    51    V    38 
P    52    V    39 
K    53    E    40 
T    54    P    41 
F    55    G    42 
Q    56    V    43 
L    57    I    44 
Q    58    G    45 
L    59    C    46 
L    60    D    47 
S    61    A    48 
P    62    T    49 
S    63    T    50 
S    64    S    51 
S    65    T    52 
V    66    D    53 
V    67    V    54 
P    68    T    55 
A    69    A    56 
F    70    D    57 
N    71    A    58 
T    72    A    59 
G    73    G    60 
T    74    K    61 
I    75    I    62 
T    76    T    63 
Q    77    A    64 
V    78    Q    65 
I    79    L    66 
K    80    K    67 
V    81    V    68 
L    82    H    69 
N    83    S    70 
P    84    S    71 
Q    85    F    72 
K    86    Q    73 
TER 
END 
