Inter-platform concordance of gene expression data for the prediction of chemical mode of action

Biology Direct

Table 1 Accuracies of predicting MOA’s in the adjusted test set, based on classifiers developed on gene expression sets profiled from microarray and RNASeq platforms

Platform	Classifier	Overall Acc. %	Sensitivity, Specificity
			PPARA	CAR/PXR	Control
Microarray	Ensemble	75	89,67	44,94	100,67
	SVM	58	56,59	33,73	100,44
	RF	67	89,54	22,94	100,56
	PLS+LDA	71	67,73	56,80	100,61
	PLS+RF	58	44,68	44,68	100,45
	PCA+LDA	17	0,27	0,27	67,0
	PCA+RF	33	33,33	0,53	83,16
	RPART	62	100,39	11,93	83,55
RNASeq	Ensemble	67	56,74	56,74	100,56
	SVM	58	67,54	22,81	100,45
	RF	58	67,54	22,81	100,45
	PLS+LDA	67	56,74	56,74	100,56
	PLS+RF	58	67,54	22,81	100,45
	PCA+LDA	25	33,20	0,40	50,17
	PCA+RF	20	22,19	11,25	33,16
	RPART	46	22,60	33,54	100,28