DNdisorder: predicting protein disorder using boosting and deep networks

BMC Bioinformatics

Table 2 Performance on the CASP10 dataset

Predictor	ACC		Sensitivity		Specificity		F-measure		Sw		AUC
	Value	±SE	Value	±SE	Value	±SE	Value	±SE	Value	±SE	Value	±SE
metaprdos2(340)	77.06	0.92	64.73	1.4	89.40	0.98	41.24	2.9	54.12	1.8	0.8727	0.006
PreDisorder(125)	76.86	0.67	67.19	1.7	86.34	0.94	37.50	1.5	53.73	1.3	0.8394	0.006
POODLE(216)	76.84	0.78	62.74	1.6	90.94	0.26	43.06	1.0	53.68	1.6	0.8663	0.006
PreDNdisorder	76.55	0.75	61.74	1.8	91.36	0.61	43.42	1.5	53.10	1.5	0.8642	0.006
ZHOU-SPARKS-X (413)	75.68	0.76	64.81	1.4	86.55	0.96	36.43	1.9	51.36	1.5	0.8588	0.006
DNdisorder(424)	75.19	0.71	61.92	1.4	88.46	0.29	38.02	1.1	50.39	1.4	0.8480	0.006
CSpritz(484)	75.13	1.4	66.31	1.3	83.94	2.4	33.64	3.7	50.25	2.9	0.8215	0.007
Espritz(380)	73.16	1.6	59.24	1.4	87.08	2.6	34.58	4.7	46.31	3.2	0.8457	0.006
espritz_nopsi_X	71.98	0.97	53.10	1.5	90.87	0.77	37.56	2.4	43.97	2.0	0.8145	0.007
PrDOS-CNF(369)	70.35	0.88	41.95	1.8	98,74	0.14	52.5	1.4	40.70	1.8	0.8956	0.005
biomine_dr_mixed (478)	69.17	0.68	39.95	1.4	98.40	0.11	49.10	1.3	38.34	1.4	0.8844	0.006
biomine_dr_pdb_c (288)	67.81	1.2	36.88	2.6	98.74	0.15	47.65	2.1	35.62	2.5	0.8815	0.006
iupred_short	63.26	0.70	30.68	1.5	95.84	0.25	32.34	1.2	26.52	1.4	0.6642	0.007

Results of a benchmark of DNdisorder, PreDisorder, PreDNdisorder and a number of other disorder predictors. The evaluation was performed on 95 CASP10 targets consisting of 22673 ordered residues and 1597 disordered residues. The standard error (SE) is shown for each performance measure. Not included in this assessment are 1186 residues which were marked as ‘X’ by the CASP assessors. This table also includes a number of methods which participated in the CASP10 experiment. The predictions for these methods were downloaded from the official CASP website and the group number for these methods is provided in parenthesis. All values except AUC have been scaled by a factor of 100.

ISSN: 1471-2105