UbiSite: incorporating two-layered machine learning method with substrate motifs to predict ubiquitin-conjugation site on lysines

Table 2 Performance evaluation of the investigated features in identifying ubiquitylation sites based on five-fold cross-validation

Investigated features	Sensitivity	Specificity	Accuracy	MCC
20D binary coding (20D)	65.59 %	67.09 %	66.64 %	0.303
Amino Acid Composition (AAC)	64.34 %	65.44 %	65.11 %	0.275
Amino Acid Pair Composition (AAPC)	68.70 %	70.72 %	70.11 %	0.367
Position Weight Matrix (PWM)	68.08 %	67.99 %	68.01 %	0.334
Position-Specific Scoring Matrix (PSSM)	69.46 %	70.69 %	70.32 %	0.374
Solvent-Accessible Surface Area (SASA)	64.58 %	65.47 %	65.20 %	0.278
Secondary Structure (SS)	55.20 %	60.51 %	58.91 %	0.145

ISSN: 1752-0509