Landscape and functional repertoires of long noncoding RNAs in the pan-cancer tumor microenvironment using single-nucleus total RNA sequencing

Intratumor heterogeneity (ITH) plays crucial roles in tumor progression. However, the atlas of long noncoding RNAs (lncRNAs) in the context of ITH across multiple cancer types remains largely unexplored. Here, we analyze over 800,000 cells from ten different cancer types generated from the random-primed single-nucleus total RNA sequencing and provide a systematic landscape of lncRNAs in tumor microenvironment (TME) and malignant programs. Our study employe a robust cell annotation pipeline called scAnnotation, which allows us to identify 39 distinct cell types within the pan-cancer TME. By applying stringent criteria, we identify thousands of reliable marker genes, including both mRNAs and lncRNAs. Next, we identify sets of cell type-specific lncRNA-mRNA pairs by our LncPairs algorithm. Moreover, we identify nine expression meta-programs (MPs) associated with diverse biological processes in malignant cells across multiple cancer types. MP-specific lncRNA-transcription factor (TF) regulatory networks are further constructed and key lncRNAs and regulons that exert control over MP-specific gene expression are identified. The comprehensive atlas of lncRNAs in the pan-cancer context, coupled with the bioinformatics tools tailored for the random-primed datasets, is expected to accelerate advancements in the field of lncRNA research at the single-cell resolution.