Accurate quantification of single-cell and single-nucleus RNA-seq transcripts using distinguishing flanking k-mers

Abstract

In single-cell and single-nucleus RNA sequencing, the coexistence of nascent (unprocessed) and mature (processed) mRNA poses challenges in accurate read mapping and the interpretation of count matrices. The traditional transcriptome reference, defining the region of interest in bulk RNA-seq, restricts its focus to mature mRNA transcripts. This restriction leads to two problems: reads originating outside of the region of interest are prone to mismapping within this region, and additionally, such external reads cannot be matched to specific transcript targets. Expanding the region of interest to encompass both nascent and mature mRNA transcript targets provides a more comprehensive framework for RNA-seq analysis. Here, we introduce the concept of distinguishing flanking k-mers (DFKs) to improve mapping of sequencing reads. We have developed an algorithm to identify DFKs, which serve as a sophisticated background filter, enhancing the accuracy of mRNA quantification. This dual strategy of an expanded region of interest coupled with the use of DFKs enhances the precision in quantifying both mature and nascent mRNA molecules, as well as in delineating reads of ambiguous status.