Table 1: Feature selected and identified metabolites in blood plasma samples of high and low-grade Gleason PCa using different multivariate models..

Table 3: Feature selected and identified metabolites in blood plasma samples of high and low-grade Gleason PCa using different multivariate models.

Feature selection methods	Random forest	PLS-DA	OPLS-DA	OL
Metabolites panels	Phosphocreatine	Histidine	Histidine	Phosphocreatine
	Choline	Choline	Choline	Choline
	Histidine	Phosphocreatine	Phosphocreatine	Histidine
	Lactate	Lactate	Lactate	Lactate
	Valine	Tyrosine	Tyrosine	Tyrosine
	Creatine	Alanine	Glucose
	Tyrosine	Glutamate	Betaine
	Taurine	Betaine	Taurine
	Alanine	Taurine	Alanine
	Glucose	Phenylalanine	Isoleucine

PLS-DA: Partial Least Squares Regression Discriminant Analysis; OPLS-DA: Orthogonal Partial Least Squares Regression Discriminant Analysis; OL: Overlapping Set Panel; Random Forest Features Ranked by Their Contributions to Classification Accuracy (Mean Decrease Accuracy)