Table 1.

Confusion matrix using a PWM based classification approach

		Actual Class I	Actual Class II
Real data		102	102
Predicted Class I	89	69	20
Predicted Class II	115	33	82
Shuffled data		102	102
Predicted Class I	100	59	41
Predicted Class II	104	43	61

Class I: well-performing templates. Class II: poorly performing templates. Values represent the number of template sequences in each set.