Prediction and classification of ncRNAs using structural information

BMC Genomics

Table 1 SVM based highest prediction performances (on the basis of MCC) of different composition approaches for the discrimination between non-coding and coding RNAs

Approach	Main dataset				Independent dataset				CONC dataset
	SN	SP	ACC	MCC	SN	SP	ACC	MCC	SN	SP	ACC	MCC
MNC	56.67	87.13	75.08	0.47	63.95	81.15	74.34	0.46	50.19	90.48	77.48	0.46
DNC	97.13	94.27	95.40	0.91	95.93	95.08	95.42	0.91	81.54	91.15	88.04	0.73
TNC	98.90	99.04	98.98	0.98	98.65	98.79	98.74	0.97	86.33	95.39	92.47	0.83
TTNC	99.50	99.35	99.41	0.99	99.29	99.05	99.15	0.98	89.51	94.98	93.22	0.85
PNC	98.62	98.55	98.58	0.97	98.78	98.29	98.49	0.97	88.43	95.64	93.31	0.85
Hybrid	99.46	99.46	99.46	0.99	99.11	99.18	99.15	0.98	89.10	96.29	93.97	0.86

The MNC, DNA, TNC, TTNC and PNC approaches are mono-, di-, tri-, tetra- and penta-nucleotide compositions respectively. Hybrid is a combined approach based prediction of all predicated SVM scores of MNC, DNC, TNC, TTNC and PNC approaches.

ISSN: 1471-2164