Information Technology Reference
In-Depth Information
Table 2.
SP values given by several methods on the BAliBase v.1.0 benchmark
(http://bips.u-strasbg.fr/fr/Products/Databases/BAliBASE/) for multiple sequence
alignment
Aligner
Ref 1(82) Ref 2(23) Ref 3(12) Ref 4(12) Ref 5(12) Overall(141)
Hybrid CSA
80
.
7
88
.
6
77
.
4
70
.
2
82
.
0
79
.
7
DIALIGN [20]
77
.
7
38
.
4
28
.
8
85
.
2
83
.
6
62
.
7
CLUSTALX [12]
85
.
3
58
.
3
40
.
8
36
.
0
70
.
6
58
.
2
82
.
2
42
.
8
33
.
3
59
.
1
63
.
8
56
.
2
PILEUP8 [11]
80
.
1
37
.
1
34
.
0
70
.
4
57
.
2
55
.
7
ML PIMA [17]
PRRP [22]
86
.
6
54
.
0
48
.
7
13
.
4
70
.
0
54
.
5
SAGA [25]
70
.
3
58
.
6
46
.
2
28
.
8
64
.
1
53
.
6
SB PIMA [17]
81
.
1
37
.
9
24
.
4
72
.
6
50
.
7
53
.
3
MULTALIGN [10]
82
.
3
51
.
6
27
.
6
29
.
2
62
.
7
50
.
6
Table 3.
Alignment accuracies given by several methods on the the BAliBASE v.2.0
benchmark (http://bips.u-strasbg.fr/fr/Products/Databases/BAliBASE2/) for multi-
ple sequence alignment [15]
Aligner
Ref 1(82) Ref 2(23) Ref 3(12) Ref 4(12) Ref 5(12) Overall(141)
SP
CS
SP
CS
SP
CS
SP
CS
SP
CS
SP
CS
SPEM [15]
90
.
8
83
.
9 93
.
4 57
.
3 81
.
4 56
.
9
97
.
4 90
.
8
97
.
4
92
.
3 91
.
5
78
.
6
MUSCLE [24]
90
.
3
84
.
7
64
.
4 60
.
9 82
.
2 61
.
9 91
.
8 74
.
8
98
.
1
92
.
1 91
.
0
78
.
7
ProbCons [16] 90
.
0 83
.
9
94
.
0 62
.
6 82
.
3 63
.
1
90
.
9 73
.
6
98
.
1
91
.
7 90
.
8
78
.
4
T-Coffee [14]
86
.
8 80
.
0 93
.
9 58
.
5 76
.
7 54
.
8 92
.
1 76
.
8 94
.
6 86
.
1 88
.
2
74
.
6
PRALINE [27] 90
.
4 83
.
9
94
.
0
61
.
0 76
.
4 55
.
8 79
.
9 53
.
9 81
.
8 68
.
6 88
.
2
73
.
9
ClustalW [13] 85
.
8 78
.
3 93
.
3 59
.
3 72
.
3 48
.
1 83
.
4 62
.
3 85
.
8 63
.
4 85
.
7
70
.
0
Hybrid CSA
82
.
7 65
.
3 91
.
9 41
.
3 78
.
6 36
.
2 70
.
5 31
.
9 83
.
6 56
.
9 81
.
4
46
.
3
reference alignments with more than 2100 sequences. The three new references
contain 26 protein families with 12 distinct repeat types, 8 transmembrane fam-
ilies and 5 families with inverted domains.
Table 2 shows the average SP score obtained by the described alignment tools
on every instance set of BAliBASE v.1.0. The values refer to the Sum of Pairs
score, calculated by the ”
baliscore.c
” program. As it can be seen in the table,
Hybrid CSA performs well on the Ref 2 and Ref 3 sets. The values obtained aid
to raise the overall score, which is higher compared to the results published by
the Bioinformatic platform of Strasbourg
1
.
Table 3 shows the average SP and Column Score (CS) values obtained by
the compared tools on every group of instances belonging to the BAliBASE
v.2.0 database. The Column Score is defined as the number of correctly aligned
columns present in the generated alignments, divided by the total number of
aligned columns in the core blocks of the reference alignment.
1
http://bips.u-strasbg.fr/fr/Products/Databases/BAliBASE/
