Table 2.

Data statistics after using CD-HIT.

Sequence identity	Training data set (6259)		Testing data set (35494)
	Positive	Negative	Positive	Negative
100% (original)	23949	228441	110695	1217977
90%	21621	196808	38739	325640
80%	21165	179691	36647	284713
70%	20709	165560	35165	255134
60%	18588	115296	29810	162044
50%	10216	34428	14210	41700
40%	2658	5532	3267	6214