Table 1. Ventral tegmental area model performance.

Cell sex classification of the ventral tegmental area (VTA) testing partition was performed using two non-ML and four ML models. Model performance was assessed by accuracy and AUC-ROC. Model: Name of classification model; Predictors: set of predictor variables in model; Training time: total model training time in hours (hr), minutes (min), and seconds (s); Overall: performance measured across all cells; Neuronal: performance measured across only neuronal cell types; Non-neuronal: performance measured across only non-neuronal cell types; AUC-ROC: area under the curve of the receiver operating characteristic curve; Accuracy: proportion of correct classifications out of all classifications.

			Overall		Neuronal		Non-Neuronal
Model	Predictors	Training time (hr:min:s)	AUC-ROC	Accuracy	AUC-ROC	Accuracy	AUC-ROC	Accuracy
Xist	Xist (ENSRNOG00000065796)	N/A	0.692	0.689	0.768	0.799	0.670	0.648
Chr Y	Chromosome Y genes	N/A	0.892	0.892	0.965	0.954	0.861	0.869
Logistic Regression (LR)	285 Selected features	00:00:29	0.978	0.917	0.994	0.971	0.968	0.898
Support Vector Machine (SVM)	285 Selected features	26:12:51	0.978	0.920	0.995	0.970	0.969	0.902
Random Forest (RF)	285 Selected features	14:47:29	0.976	0.913	0.996	0.970	0.963	0.893
Multi-Layer-Perceptron (MLP)	285 Selected features	07:31:27	0.966	0.907	0.988	0.948	0.955	0.892