Abstract
Single-cell sequencing enables genome- and epigenome-wide profiling of thousands of individual cells, offering unprecedented biological insights. However, technical noise and batch effects obscure high-resolution structures, hindering rare-cell-type detection and cross-dataset comparisons. To comprehensively address these challenges, this study upgrades RECODE, a high-dimensional statistics-based tool for technical noise reduction in single-cell RNA sequencing (RNA-seq), to include a function called iRECODE, which simultaneously reduces technical and batch noise. Further, RECODE's applicability is extended to diverse single-cell modalities, including single-cell high-throughput chromosome conformation capture (Hi-C) and spatial transcriptomics. Recent improvements in the algorithm have substantially enhanced both accuracy and computational efficiency. The RECODE platform thus provides a robust and versatile solution for noise mitigation, enabling more accurate downstream analyses across transcriptomic, epigenomic, and spatial domains.