Predicting sumoylation sites using support vector machines based on various sequence features, conformational flexibility and disorder

Table 4 Effect of conformational flexibility and disorder on prediction performance.

	Self-consistency				5-fold Cross Validation
Information	Acc	Sp	Sn	MCC	Acc	Sp	Sn	MCC
All Features	0.97	0.98	0.76	0.68	0.97	0.98	0.73	0.66
without Flexibility	0.97	0.98	0.75	0.67	0.97	0.98	0.72	0.66
without Disorder	0.97	0.98	0.75	0.67	0.97	0.98	0.72	0.66
without Flexibility & Disorder	0.97	0.98	0.74	0.67	0.97	0.98	0.72	0.66

Conformational flexibility and disorder features ('DisorderBinary', 'DisorderReal') were eliminated from dataset one by one and together. Self-consistency and 5-fold cross validation tests were performed. An average of 25 repeats are reported for 5-fold cross validation tests. Standard deviations were less than 0.01, so they were not reported.

ISSN: 1471-2164