Contact prediction in protein modeling: Scoring, folding and refinement of coarse-grained models

BMC Structural Biology

Table 2 Results of the contact-based ranking of Kolinski-Bujnicki's models for NF and FR/A CASP6 targets.

Set of contact data	Number of predicted contacts	Accuracy δ = 0 [%]	Coverage δ = 0 [%]	Accuracy δ = 2 [%]	False contacts (δ > 5) [%]	ΔRMSD [Å]	Spearman corr. coeff. (GDT-TS)	Spearman corr. coeff. (RMSD)
N/2 top-scoring contacts from each of the best two predictors ^(a)	N	19.58	12.91	49.37	19.59	1.069	-0.300	0.329
N/2 top-scoring contacts from each of the best three predictors^(b)	1.5 N	17.21	15.22	46.90	21.13	1.453	-0.321	0.275
Consensus of the whole data from the best three predictors	N/2	23.94	9.12	53.17	17.93	1.393	-0.325	0.344
	N	18.90	14.44	49.17	20.35	1.387	-0.333	0.340
	1.5 N	15.53	17.83	46.60	22.15	1.453	-0.329	0.288
Consensus of the whole data from the best five predictors^(c)	N/2	23.78	8.97	52.47	17.67	1.272	-0.196	0.360
	N	19.78	15.07	51.01	20.04	1.498	-0.350	0.348
	1.5 N	16.64	18.91	47.70	22.11	1.443	-0.338	0.400
Consensus of the whole data from all nine predictors	N/2	24.98	9.50	51.69	20.56	1.252	-0.238	0.432
	N	20.63	15.74	50.58	21.85	1.365	-0.333	0.400
	1.5 N	18.08	20.37	49.35	23.39	1.322	-0.342	0.440

^(a) Predictors: Baker and PROFcon
^(b) Predictors: Baker, PROFcon, GPCPRED
^(c) Predictors: Baker, PROFcon, GPCPRED, Karypis, SAM-T04
Spearman correlation coefficients together with the average difference of Cα RMSD of the best model and the first ranked model (ΔRMSD) were computed for each set of contact data. The average ΔRMSD of the Kolinski-Bujnicki group was 1.484 and the Spearman correlation coefficient: 0.213 (RMSD) and -0.138 (GDT-score). These values were improved for every set of contact data. Apart from the mean accuracy and the mean coverage of contact data, averaged over all targets, we also present the percentage of semi-accurate contacts (shifted with respect to the real by at most two residues) and totally false contacts (shifted by more than five residues).

ISSN: 1472-6807