Table 3. Performance on large-scale prediction.

Test dataset	Sn	Sp	MCC	Ac	Sw	AUC
S3845a	0.80	0.88	0.45	0.88	0.69	0.91
S1508b	0.77	0.89	0.45	0.88	0.66	0.91
1-domainc	0.80	0.91	0.51	0.90	0.70	0.92
2-domainc	0.77	0.89	0.46	0.89	0.66	0.91
m-domainc	0.71	0.86	0.35	0.85	0.57	0.88

^a:ten-fold cross-validation of 3845 entries.

^b:independent test (1508 entries) by training on the entire S3845.

^c:sequences inS1508.