Efficient pairwise RNA structure prediction using probabilistic alignment constraints in Dynalign

BMC Bioinformatics

Table 11 Structural prediction accuracy statistics for the methods benchmarked over 2000 random 5S RNA selections when scored using an exact match criterion.

		Percent sequence similarity
			40–60	60–80	80–100	0–100
Dynalign new constraint	Sens	0.850	0.870	0.876	0.869	0.871
	PPV	0.790	0.791	0.782	0.743	0.785
Dynalign M constraint	Sens	0.847	0.869	0.876	0.869	0.870
	PPV	0.789	0.789	0.778	0.744	0.782
FOLDALIGN	Sens	0.684	0.721	0.772	0.501	0.724
	PPV	0.724	0.750	0.789	0.562	0.752
StemLoc	Sens	0.253	0.631	0.858	0.711	0.698
	PPV	0.603	0.751	0.854	0.738	0.786
Consan	Sens	0.602	0.743	0.882	0.759	0.786
	PPV	0.598	0.704	0.768	0.608	0.717
Single Prediction	Sens	0.655	0.695	0.734	0.744	0.709
	PPV	0.579	0.606	0.641	0.632	0.618

Results are summarized for sequence similarity values ranging from 20% through 100% in steps of 20% and for the overall data set (0 – 100). Dynalign new constraint refers to Dynalign with probabilistic alignment constraints proposed here. Software version numbers and parameters for the algorithms are described in Section 7.6.

ISSN: 1471-2105