Correcting Cross-Cell UMI Bias for Improved Single-Cell Transcriptomics

Accurate quantification of mRNA at the single-cell level is crucial for understanding gene expression dynamics, cellular diversity, and disease mechanisms.Traditional normalization methods for single-cell RNA sequencing (scRNA-seq) typically measure relative RNA abundance, which limits their ability to provide absolute quantification and does not account for variations in total RNA content across cells.To address this limitation, we present the LOESS regression-corrected UMI count (LC-UMI), a method that adjusts UMI counts to correct cross-cell amplification variability, enabling more accurate gene expression quantification at the single-cell level.We demonstrate that LC-UMI accurately captures differences in RNA content across cell types.At the single-cell level, total LC-UMIs strongly correlate with transcriptionrelated pathways, such as RNA polymerase complex and mitochondrial oxidative phosphorylation.In downstream analyses, including cell clustering, differential gene expression, and pseudotime analysis, log(LC-UMI+1) delivers results comparable to conventional normalization methods.LC-UMI is implemented in the R package LcUMI, which is freely available for any UMI-based scRNA-seq dataset.These findings underscore the potential of LC-UMI to enhance the precision and reliability of single-cell transcriptomics, providing a valuable metric for future research in single-cell gene expression.