Predicting human splicing branchpoints by combining sequence-derived features and multi-label learning methods

BMC Bioinformatics

Table 1 The performance of multi-label learning methods based on different features

Method	Feature	Recall	Precision	ACC	F	AUC	AUPR
PLS	Markov	0.508	0.478	0.961	0.473	0.879	0.476
	PWM	0.521	0.454	0.958	0.465	0.868	0.455
	DN	0.534	0.437	0.957	0.461	0.877	0.461
	SP	0.545	0.455	0.958	0.477	0.874	0.483
	PPT	0.574	0.098	0.787	0.170	0.698	0.103
CCA	Markov	0.529	0.461	0.959	0.472	0.880	0.476
	PWM	0.521	0.453	0.958	0.465	0.868	0.455
	DN	0.566	0.423	0.955	0.466	0.883	0.468
	SP	0.533	0.466	0.960	0.477	0.878	0.485
	PPT	0.488	0.118	0.844	0.182	0.703	0.114
LSCCA	Markov	0.502	0.501	0.963	0.482	0.882	0.486
	PWM	0.516	0.471	0.960	0.472	0.871	0.467
	DN	0.546	0.442	0.957	0.469	0.883	0.453
	SP	0.513	0.513	0.963	0.494	0.882	0.487
	PPT	0.472	0.085	0.790	0.129	0.690	0.086

ISSN: 1471-2105