Table 1.
Performance evaluation of the new version of our method HMM-ModE on ‘gold standard’ dataset and comparison with HMM-ModE/HMMER2 [9]
| SuperFamily | Family | # Seq |
HMM-d |
HMM-ModE with HMMER3 (HMMER2) |
||
|---|---|---|---|---|---|---|
| Se | Sp | Se | Sp | |||
| AminoHydrolase |
AMP deaminase |
28 |
1 |
0.90 |
1(1) |
1(1) |
| urease |
100 |
1 |
0.50 |
1(1) |
1(1) |
|
| D-hydantoinase |
10 |
1 |
0.05 |
1(1) |
1(1) |
|
| Dihydroorotase 2 |
13 |
1 |
0.14 |
1(1) |
1(1) |
|
| Guanine deaminase |
11 |
1 |
0.22 |
1(1) |
1(1) |
|
| Adenosine deaminase |
10 |
1 |
0.35 |
1(1) |
1(1) |
|
| Enolase |
Enolase |
215 |
1 |
0.88 |
1(1) |
1(1) |
| Glucarate dehydratase |
26 |
1 |
0.20 |
1(1) |
1(1) |
|
| Muconate cycloisomerase |
14 |
1 |
0.09 |
1(1) |
1(1) |
|
| Chloromuconate cycloisomerase |
10 |
1 |
0.05 |
1(1) |
1(1) |
|
| Crotonase |
Enoyl-CoA hydratase |
54 |
1 |
0.23 |
1(1) |
1(1) |
| Histone acetyltransferase |
11 |
11 |
0.09 |
1(1) |
1(1) |
|
| Haloacid Dehydrogenase |
P-type atpase |
91 |
1 |
0.77 |
1(1) |
1(1) |
| Vicinal Oxygen chelate | Catechol 2,3-dioxygenase |
32 |
1 |
0.28 |
1(1) |
0.88 (0.88) |
| 4-Hydroxyphenylpyruvate dioxygenase |
26 |
1 |
0.60 |
1(1) |
1(1) |
|
| 2,3-Dihydroxybiphenyl dioxygenase |
23 |
1 |
0.20 |
0.95 (1) |
0.71(0.63) |
|
| Glyoxalase 1 | 12 | 1 | 0.20 | 1(1) | 1(1) | |
HMM-d stands for default HMMER and # Seq stands for number of sequences in each family. Se is Sensitivity and Sp is Specificity. Sn is the sensitivity value calculated as TP/(TP + FN), and Sp is the specificity calculated as TP/(TP + FP). As shown here, there is a significant improvement in the specificity of the predictions with small compromise in the respective sensitivity using HMM-ModE.
The corresponding values of Se and Sp using HMM-ModE with HMMER2 is written in parenthesis.