Identification of putative domain linkers by a neural network – application to a large sequence database

BMC Bioinformatics

Table 1

Sequence regions detected	No. of sequences^a	No. of sequence regions^b	No. of residues^c	% residues^d
All	101602		37315215	100.00
PDB	38470	410090	10210325	27.36
CDD	64349	124888	16207467	43.43
Low-complexity regions (45, 3.4, 3.75)^e	48641	70373	8474412	22.71
Low-complexity regions (45, 2.9, 3.2)	6735	8539	803001	2.15
Low-complexity regions (45, 2.6, 2.9)	3208	3970	359227	0.96
Low-complexity regions (45, 2.45, 2.75)	2340	2786	250796	0.67
Putative domain linkers (0.90)^f	14239	20876	1051607	2.82
Putative domain linkers (0.91)	12670	18193	953097	2.55
Putative domain linkers (0.92)	11160	15620	856149	2.29
Putative domain linkers (0.93)	9554	13053	752119	2.02
Putative domain linkers (0.94)	7977	10591	644472	1.73
Putative domain linkers (0.95)	6387	8133	529884	1.42
Putative domain linkers (0.96)	4819	5892	415150	1.11
Putative domain linkers (0.97)	3099	3592	281009	0.75
Putative domain linkers (0.98)	1326	1469	128455	0.34
Low-complexity regions (45, 2.9, 3.2) + Putative domain linkers (0.95)^g	10364	13946	1139983	3.06

Statistics of SWISSPROT sequences. ^a Number of SWISSPROT sequences that contained the detected sequence regions. ^b Number of sequence regions detected in the SWISSPROT sequences. ^c Total number of residues in the detected sequence regions. ^d Percentage of residues in the detected regions relative to all of the residues in the SWISSPROT sequences. ^e The values of the three parameters used for the SEG program, namely, the trigger window, the trigger and extension complexities are listed in the parentheses. ^f The cutoff parameter used for our neural network is indicated in the parentheses. ^g Predictions obtained by merging putative domain linkers and the low-complexity regions.

ISSN: 1471-2105