LBSizeCleav: improved support vector machine (SVM)-based prediction of Dicer cleavage sites using loop/bulge length

Bao, Yu; Hayashida, Morihiro; Akutsu, Tatsuya

doi:10.1186/s12859-016-1353-6

Table 2 Results on average specificity, sensitivity, accuracy, and MCC for both 5p and 3p arms by five-fold cross-validation using PHDCleav and LBSizeCleav (k=1,⋯,5) with window sizes 8,10,12,14 based on sequences and secondary structures predicted by quikfold server

From: LBSizeCleav: improved support vector machine (SVM)-based prediction of Dicer cleavage sites using loop/bulge length

Method	Window size	5p arm				3p arm
		Sn	Sp	Ac	MCC	Sn	Sp	Ac	MCC
PHDCleav (sequence)	8	0.602	0.503	0.552	0.105	0.662	0.625	0.644	0.287
	10	0.541	0.573	0.557	0.115	0.661	0.642	0.652	0.303
	12	0.560	0.555	0.557	0.115	0.660	0.656	0.658	0.316
	14	0.539	0.572	0.555	0.111	0.654	0.702	0.678	0.356
PHDCleav (structure)	8	0.753	0.814	0.784	0.568	0.670	0.661	0.665	0.330
	10	0.784	0.827	0.806	0.612	0.702	0.719	0.710	0.421
	12	0.790	0.842	0.816	0.633	0.739	0.764	0.752	0.503
	14	0.799	0.857	0.828	0.657	0.779	0.783	0.781	0.562
Extended PHDCleav	8	0.750	0.798	0.774	0.548	0.652	0.716	0.684	0.369
	10	0.779	0.827	0.803	0.607	0.674	0.783	0.729	0.460
	12	0.809	0.845	0.827	0.654	0.714	0.790	0.752	0.506
	14	0.813	0.868	0.840	0.682	0.781	0.801	0.791	0.582
LBSizeCleav (k=1)	8	0.668	0.924	0.796	0.612	0.630	0.684	0.657	0.315
	10	0.709	0.947	0.828	0.675	0.651	0.776	0.713	0.430
	12	0.774	0.945	0.859	0.730	0.686	0.847	0.766	0.540
	14	0.808	0.933	0.871	0.747	0.758	0.874	0.816	0.637
LBSizeCleav (k=2)	8	0.662	0.954	0.808	0.645	0.626	0.723	0.674	0.351
	10	0.725	0.946	0.835	0.688	0.642	0.806	0.724	0.455
	12	0.784	0.938	0.861	0.731	0.665	0.882	0.773	0.560
	14	0.820	0.925	0.872	0.749	0.734	0.916	0.825	0.661
LBSizeCleav (k=3)	8	0.692	0.949	0.821	0.664	0.619	0.735	0.677	0.356
	10	0.752	0.941	0.846	0.706	0.618	0.822	0.720	0.450
	12	0.803	0.932	0.867	0.741	0.635	0.914	0.774	0.571
	14	0.825	0.912	0.869	0.740	0.719	0.942	0.830	0.678
LBSizeCleav (k=4)	8	0.695	0.949	0.822	0.667	0.614	0.736	0.675	0.353
	10	0.767	0.938	0.853	0.716	0.621	0.835	0.728	0.467
	12	0.815	0.927	0.871	0.747	0.639	0.912	0.776	0.573
	14	0.835	0.909	0.872	0.746	0.723	0.924	0.823	0.660
LBSizeCleav (k=5)	8	0.700	0.947	0.824	0.668	0.594	0.771	0.682	0.371
	10	0.772	0.936	0.854	0.717	0.578	0.862	0.720	0.459
	12	0.821	0.924	0.872	0.749	0.634	0.921	0.777	0.579
	14	0.838	0.909	0.874	0.749	0.724	0.932	0.828	0.671

Sn, Sp, Ac, and MCC denote sensitivity, specificity, accuracy, and Matthews correlation coefficient, respectively

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com

BMC Bioinformatics

Contact us