Table 1

Results of the small-scale evaluation.

Factor
Sites
Method
Model
Distinct sequences identified
Target
True positives
False positives
False positive ratio

E2F
27
HMMER
M00050
27
27
36
8
18.20%



T05206
27
27
27
4
12.90%


Match
V$E2F_02
27
27
36
7
16.30%


Patser
V$E2F_02
27
27
36
8
18.20%


LMM
V$E2F_02
18
27
18
3
n/a


ScanACE
M00050
12
27
12
3
n/a
ER
17
HMMER
M00959
16
16
22
7
24.10%



T00258
17
16
16
2
11.10%


Match
V$ER_Q6_02
17
16
24
7
22.60%


Patser
V$ER_Q6_02
17
16
24
8
25.00%


LMM
V$ER_Q6
15
16
15
0
0.00%


ScanACE
M00959
8
16
11
1
n/a
GR
7
HMMER
M00921
7
7
10
2
16.70%



T05076
7
7
7
1
12.50%


Match
V$GR_Q6_01
7
7
9
3
25.00%


Patser
V$GR_Q6_01
7
7
9
7
43.70%


LMM
V$GR_Q6
6
7
6
1
14.30%


ScanACE
M00921
4
7
4
1
n/a
HNF-1
18
HMMER
M00790
18
18
19
0
0.00%



T01211
18
18
18
0
0.00%


Match
V$HNF1_Q6
18
18
22
3
12.00%


Patser
V$HNF1_Q6
18
18
29
1
3.30%


LMM
V$HNF1_01
16
18
16
0
0.00%


ScanACE
M00790
11
18
11
0
n/a
HNF-3
10
HMMER
M00724
10
10
10
1
9.10%



T01049
10
10
10
2
16.70%


Match
V$HNF3ALPHA_Q6
10
10
12
4
25.00%


Patser
V$HNF3ALPHA_Q6
10
10
10
1
9.10%


LMM
V$HNF3ALPHA_Q6
9
10
9
4
30.80%


ScanACE
M00724
8
10
8
0
0.00%
HNF-4
10
HMMER
M00638
9
9
9
2
18.20%



T00372
10
9
9
0
0.00%


Match
V$HNF4ALPHA_Q6
9
9
9
2
18.20%


Patser
V$HNF4ALPHA_Q6
10
9
9
2
18.20%


LMM
V$HNF4ALPHA_Q6
7
9
7
0
0.00%


ScanACE
M00638
3
9
3
0
n/a

The "Sites" column contains the number of sequences containing experimentally validated binding sites provided as input. "Target" represents the number of binding sites to be retrieved by a method to be considered successful, and "Distinct sequences identified" is the number of distinct sequences in which at least one true positive was detected. Because of partially overlapping hits, the actual number of true positives reported may be higher than the true number of sites. Not all matrices tested were available in the LMM matrix library; for those cases, the results obtained using the closest available LMM matrix are displayed in italics.

Marinescu et al. BMC Bioinformatics 2005 6:79   doi:10.1186/1471-2105-6-79