Significant improvement of miRNA target prediction accuracy in large datasets using meta-strategy based on comprehensive voting and artificial neural networks

Table 2 Numbers of samples in each of the eleven newly-designed datasets

Dataset	Associated individual predictors	No. of original samples	No. of non-redundant samples	No. of positive samples	No. of negative samples	No. of miRNAs	No. of mRNAs
D4	miRanda, MiRDB, PITA, TargetScan	45,517	22,446	1844	20,602	211	6747
D3–1	miRanda, MiRDB, PITA	29,486	9271	1339	7932	212	4619
D3–2	miRanda, MiRDB, TargetScan	7584	2478	198	2280	201	1097
D3–3	miRanda, PITA, TargetScan	107,813	5529	1220	4309	205	3541
D3–4	MiRDB, PITA, TargetScan	66,384	2984	864	2120	323	1946
D2–1	miRanda, MiRDB	5269	892	199	693	186	641
D2–2	miRanda, PITA	216,923	974	457	517	202	883
D2–3	miRanda, TargetScan	32,566	429	151	278	162	342
D2–4	MiRDB, PITA	29,531	810	455	355	165	645
D2–5	MiRDB, TargetScan	256,784	430	174	256	259	337
D2–6	PITA, TargetScan	384,944	363	179	184	175	288

ISSN: 1471-2164