Accurate classification of membrane protein types based on sequence and evolutionary information using deep learning

Table 5 Comparison of the Se, Sp, ACC and MCC between four traditional models and sequence information model on Dataset 1

Membrane protein types	Index	AAC	DipC	TipC	PseAAC	Our model
	Se	0.7928	0.7095	0.5225	0.8221	0.8356
Type I	Sp	0.9537	0.9671	0.9805	0.9568	0.9941
	ACC	0.9372	0.9407	0.9335	0.9430	0.9778
	MCC	0.6898	0.6772	0.5936	0.7192	0.8751
	Se	0.2692	0.0641	0.0256	0.2949	0.8077
Type II	Sp	0.9946	0.9993	1.0	0.9951	0.9871
	ACC	0.9815	0.9825	0.9825	0.9825	0.9838
	MCC	0.3499	0.1964	0.1587	0.3845	0.6492
	Se	0.0	0.0	0.0	0.0	0.1667
Type III	Sp	1.0	1.0	1.0	1.0	0.9998
	ACC	0.9986	0.9986	0.9986	0.9986	0.9986
	MCC	0.0	0.0	0.0	0.0	0.2881
	Se	0.0833	0.0	0.0	0.0	0.6667
Type IV	Sp	1.0	1.0	1.0	1.0	0.9984
	ACC	0.9975	0.9972	0.9972	0.9972	0.9986
	MCC	0.2883	0.0	0.0	0.0	0.5950
	Se	0.9069	0.9587	0.9816	0.9029	0.9461
Mutipass	Sp	0.8661	0.7004	0.4766	0.8858	0.9092
	ACC	0.8968	0.8950	0.8571	0.8987	0.9370
	MCC	0.7390	0.7051	0.5847	0.7479	0.8355
	Se	0.1579	0.0789	0.0263	0.1579	0.1579
Lipid-chain-anchored	Sp	0.9988	1.0	1.0	0.9993	0.9979
	ACC	0.9915	0.9919	0.9915	0.9919	0.9905
	MCC	0.2904	0.2798	0.1615	0.3219	0.2473
	Se	0.3696	0.0435	0.0435	0.3478	0.8478
GPI-anchored	Sp	0.9974	1.0	1.0	0.9960	0.9951
	ACC	0.9908	0.9898	0.9898	0.9892	0.9935
	MCC	0.4694	0.2075	0.2075	0.4054	0.7393
	Se	0.7342	0.6261	0.4662	0.8041	0.8806
Peripheral	Sp	0.9262	0.9617	0.9874	0.9262	0.9609
	ACC	0.9065	0.9273	0.9340	0.9137	0.9527
	MCC	0.5747	0.5981	0.5835	0.6226	0.7708

ISSN: 1471-2105