Table 5. Performance comparison among various types of sequence features and methods for the novel-germline prediction.

Method	Sensitivity	Specificity	Accuracy (%)
Random Forests (AAC)	0.626	0.581	60.65
Random Forests (DPC)	0.785	0.640	72.22
Random Forests (PPs)	0.675	0.484	59.26
Random Forests (AAC+DPC)	0.768	0.634	71.06
Random Forests (AAC+PPs)	0.671	0.522	60.65
Random Forests (DPC+PPs)	0.732	0.629	68.75
Random Forests (AAC+DPC+PPs)	0.699	0.602	65.74
BLAST [44]	0.411	0.645	51.16