Fig. 3.

AUC of leprosy risk RNA-Seq and RT-qPCR signatures in blood. Area Under the Curve (AUC) of risk signatures in whole blood to prospectively predict leprosy progressors within household contacts (HC). The models were built using Random Forest, were trained with 80% of the sample sets and evaluated in 20%. a) AUC of RNA-Seq 19-gene signature where 8 to 20 features/genes were automatically selected by the model from a total of 1,613 features. b) AUC of RNA-Seq 13-gene signature based on the 19-gene signature but excluding pseudogenes and long non-coding (lnc)RNA (n=6). c) AUC of reverse transcription quantitative PCR (RT-qPCR) 13-gene signature selected in the RNA-Seq signature. d) AUC of RT-qPCR 4-gene signature RISK4LEP (final RT-qPCR signature) where only genes significantly differentially expressed in the RT-qPCR were selected.