Prediction and prioritization of autism-associated long non-coding RNAs using gene expression and sequence features

Table 1 Model performance with both developmental brain gene expression features and RNA transcript sequence features

Model	Metric*	Expression features	Expression and sequence features
LR	ROC AUC	0.8289 ± 0.0030	0.8313 ± 0.0024
	PR AUC	0.7096 ± 0.0051	0.7177 ± 0.0044
	Accuracy	0.7615 ± 0.0059	0.7563 ± 0.0054
	Sensitivity	0.7377 ± 0.0071	0.7498 ± 0.0074
	Specificity	0.7719 ± 0.0087	0.7602 ± 0.0087
	MCC	0.4723 ± 0.0095	0.4697 ± 0.0079
SVM	ROC AUC	0.8232 ± 0.0049	0.8217 ± 0.0066
	PR AUC	0.7085 ± 0.0056	0.7159 ± 0.0081
	Accuracy	0.7746 ± 0.0052	0.7745 ± 0.0076
	Sensitivity	0.7111 ± 0.0115	0.7035 ± 0.0156
	Specificity	0.7999 ± 0.0094	0.8023 ± 0.0134
	MCC	0.4826 ± 0.0087	0.4789 ± 0.0123
RF	ROC AUC	0.8187 ± 0.0052	0.8258 ± 0.0053
	PR AUC	0.6908 ± 0.0075	0.7008 ± 0.0113
	Accuracy	0.7767 ± 0.0196	0.7699 ± 0.0188
	Sensitivity	0.6478 ± 0.0468	0.6981 ± 0.0497
	Specificity	0.8271 ± 0.0433	0.7986 ± 0.0401
	MCC	0.4644 ± 0.0207	0.4711 ± 0.0200

*The mean ROC AUC, PR AUC, overall accuracy, sensitivity, specificity and Matthews Correlation Coefficient (MCC) of the models from 50 repetitions of tenfold cross-validations are shown

ISSN: 1471-2105