长读长RNA测序在肺癌异构体与融合基因发现中的应用价值

Isoform expression is frequently dysregulated in lung cancer via cis-acting splice site mutations or trans-acting mutations in splicing factors such as U2AF1, RBM10, and SF3B1. While prior studies have sought to characterize the isoform landscape in cancers, the short reads of next-generation sequencing platforms, ranging in length from 50-150 bp, preclude the accurate phasing of alternative splicing events across the full transcript length (typically exceeding 1 kb). To provide a more comprehensive and accurate view of expressed isoforms in lung cancer, we performed long-read RNA-sequencing on 32 matched tumor/normal pairs of early stage resected lung adenocarcinoma, four biopsies from tumors that had progressed on targeted therapy treatment, and 14 non-small cell lung cancer cell lines. RNA was isolated from fresh frozen tissue specimens and cDNA prepared using the PacBio Kinnex full-length isoform method. PacBio HiFi data were generated per manufacturer's recommendations at the University of Washington Long Reads Sequencing Center or UC Davis DNA Technologies core. Libraries were sequenced on the PacBio Revio system to a read depth of >10M HiFi reads per sample, providing >80% saturation of known isoforms. FLAIR3 was used to identify and quantify isoforms, including novel isoforms, and phase isoform expression with somatic mutations including SNVs and insertion/deletion mutations. Analysis of alternative splicing patterns in RAS-pathway genes identified increased expression of the minor KRAS isoform, KRAS4A , in tumors compared to normal samples. Phasing of somatic variants enabled integration of KRAS mutation with KRAS isoform expression and validated the previously identified role of KRAS Q61 variants on aberrant KRAS splicing. In addition, we identified deletions predicted to inactivate tumor suppressor genes and identified novel rearrangements in clinically actionable oncogenes including EGFR . Together these data demonstrate the utility of long-read RNA sequencing for accurate and complete isoform characterization in cancer.