MOCHI enables systematic identification of candidate independent intronic transcriptional signals from short-read RNA-seq data

MOCHI identifies candidate independent intronic transcription units from strand-specific short-read RNA-seq, recovering sense-intronic long non-coding RNAs and enriching host introns with initiation signals.

• Why it matters: Standard pipelines ignore intronic reads, and prior approaches rely on annotation, crude whole-intron counting, or manual inspection, limiting systematic discovery of independently transcribed intronic RNAs.
• What they did: MOCHI combines reference-independent k-mer signal discovery with reference-guided interpretation plus multi-step filtering; it recovered annotated sense-intronic long non-coding RNAs in two simulated datasets and curated candidates across paired rRNA-depleted and poly(A)-selected libraries from ten cancer cell lines.
• The result: Candidate intronic expression is largely uncoupled from host-gene exons, genomic features resemble annotated long non-coding RNAs, and in ten TCGA matched cohorts candidates are detected in tumors with a subset linked to overall survival after adjustment for age and host-gene expression.

AI summary of the abstract · may contain errors · not medical advice · read the paper.