Advancing leprosy risk prediction through identification of a whole blood host transcriptomic biomarker signature including non-coding genes.

Leprosy is a debilitating disease that requires early detection for effective control, yet diagnosis still relies on clinical signs. Previous RNA-Seq analysis of coding genes from leprosy patients' household contacts (HHC) who developed leprosy (progressors) and those who did not, revealed a 4-gene RNA signature, RISK4LEP, that predicted leprosy 4-61 months before clinical onset (AUC: 0.86). To improve this signature, the present study included non-coding genes and applied novel Differential Gene Expression (DGE) analyses and machine learning approaches to the RNA-Seq dataset. This strategy identified significant DGE between progressors and HHC for 40 genes. Next, the 10 most significantly different expressed genes, as well as genes from the optimal 3-gene signatures, were validated by RT-qPCR in an independent cohort. This analysis confirmed the diagnostic potential to discriminate progressors from HHC for 12 genes. Moreover, RPS21 and SNHG5 genes were each significantly higher expressed in progressors compared to diagnosed leprosy patients, suggesting their temporary role during early (preclinical) leprosy. Furthermore, the optimal 3-gene signature consisted of two non-coding genes and one coding gene (SNHG5, SNHG8, C6orf48; sensitivity: 88%; specificity: 88%; AUC: 0.96). This study thus identified an improved prospective host transcriptional risk signature in blood based on non-coding genes predicting the development of leprosy.