TORC: Target-Oriented Reference Construction for Supervised Cell-Type Identification in scRNA-seq.

Cell-type identification is a crucial step in single-cell RNA-seq (scRNA-seq) data analysis, for which supervised methods are preferred due to their accuracy and efficiency. The quality of the reference data plays an important role in cell-type identification performance, but systematic strategies for selecting and reconstructing reference data remain limited. We present Target-Oriented Reference Construction (TORC), a widely applicable strategy for constructing reference data from available labeled cells given a target dataset. TORC alleviates the differences in data distribution and cell-type composition between the reference and the target. TORC combines initial supervised prediction, optional reference expansion using target cells with high-confidence predicted labels, and reference reconstruction guided by estimated cell-type compositions. Here, we provide detailed, step-by-step instructions describing the input requirements, configurable parameters, and practical considerations for applying TORC in real scRNA-seq analyses. TORC is available at https://github.com/weix21/TORC , where an example implementation using an MLP-based classifier is provided.