CFC-seq: identification of full-length capped RNAs unveil enhancer-derived transcription

Long-read sequencing has emerged as a powerful tool for uncovering novel transcripts and genes. However, existing protocols often lack confidence in identifying the transcription start site (TSS) and fail to capture non-poly(A) RNA, thereby limiting the discovery of novel genes, particularly long non-coding RNAs (lncRNAs). In this study, we introduce Cap-trap full-length cDNA sequencing (CFC-seq), a comprehensive protocol that combines Cap-trapping and poly(A)-tailing with Oxford Nanopore sequencing. This protocol enables precise identification of TSSs and full-length transcripts. Applying CFC-seq to two in vitro differentiation time courses resulted in approximately 236 million mappable reads. The transcript Start-site Aware Long-read Assembler (SALA) was developed for de novo assembling the transcript models, leading to the identification of 39,425 confident novel genes. Using this dataset, enhancer-derived ncRNAs were re-defined with longer length and more splicing activity, which were correlated with enhancer structure. Compared to enhancers with CpG islands, TATA box enhancers were shown to be more cell type specific with fewer chromatin interaction but produced longer and more stable polyadenylated RNA. A significant proportion of these TATA box-derived eRNAs originated from LTR transposable elements. Overall, this study systematically annotated ∼24,000 novel eRNA genes and correlated their transcription properties with enhancer structure.