Table 3.

Performance comparison with other predictors on training and independent datasets.

Residue	Methods	10-Fold Cross-Validation Test (P.ELM)				Independent Dataset Test (PPA)
		Sn	Sp	MCC	AUC	Sn	Sp	MCC	AUC
S	GPS 2.1	33.07	93.29	0.201	0.741	22.20	95.26	0.135	0.670
	NetPhos	34.14	86.73	0.123	0.702	28.55	87.23	0.081	0.643
	PPRED	32.27	91.64	0.169	0.751	21.32	94.00	0.107	0.676
	Musite	41.37	93.66	0.249	0.807	28.60	95.21	0.182	0.726
	PhosphoSVM	44.43	94.04	0.298	0.841	34.01	95.90	0.237	0.776
	SKIPHOS	78.50	74.90	0.521	0.845	46.20	68.60	0.265	0.691
	DeepPhos	81.81	75.30	0.572	0.859	66.43	75.89	0.425	0.775
	TransPhos	80.56	75.80	0.564	0.858	67.16	75.89	0.432	0.787
T	GPS 2.1	38.10	92.30	0.201	0.695	13.48	94.51	0.067	0.572
	NetPhos	34.32	83.65	0.090	0.655	27.02	80.66	0.038	0.554
	PPRED	30.31	90.99	0.134	0.726	26.43	83.51	0.052	0.578
	Musite	33.84	94.76	0.221	0.785	15.56	95.36	0.098	0.622
	PhosphoSVM	37.31	94.99	0.251	0.818	21.79	93.41	0.115	0.665
	SKIPHOS	74.40	78.80	0.547	0.844	65.80	58.60	0.197	0.643
	DeepPhos	77.63	73.58	0.512	0.826	46.02	76.04	0.231	0.674
	TransPhos	76.54	74.70	0.512	0.834	47.32	76.22	0.246	0.672
Y	GPS 2.1	34.49	78.86	0.083	0.611	47.93	60.83	0.043	0.552
	NetPhos	34.66	84.45	0.132	0.653	63.91	46.10	0.048	0.554
	PPRED	43.04	82.65	0.169	0.702	42.01	65.08	0.064	0.539
	Musite	38.42	86.74	0.182	0.720	28.85	81.71	0.064	0.587
	PhosphoSVM	41.92	87.34	0.209	0.738	28.55	84.39	0.084	0.595
	SKIPHOS	71.10	69.10	0.396	0.700	65.80	58.60	0.197	0.634
	DeepPhos	69.01	64.22	0.332	0.714	49.93	66.37	0.165	0.621
	TransPhos	61.99	65.11	0.271	0.695	38.52	72.30	0.115	0.601

The left half is the result of 10-fold cross-validation on the training dataset, and the right half is the result on the independent test set. Sensitivity (Sn), Specificity (Sp), Matthews correlation coefficient (MCC) and Area under curve (AUC) were calculated to measure the performance of models. Data in bold indicates that the model performs best for that evaluation metric.