Novel splicing and open reading frames revealed by long-read direct RNA sequencing of adenovirus transcripts

Alexander M. Price; Katharina E. Hayer; Daniel P. Depledge; Angus C. Wilson; Matthew D. Weitzman

Fractions of these three samples were prepared into standard strand specific Illumina RNA-seq 169 libraries using the polyadenylated mRNA fraction. The same RNA samples were then poly(A) 170 purified before submitting to direct RNA sequencing (dRNA-seq) on an Oxford Nanopore Using full-length RNAs, we were able to detect novel splice sites producing canonical ORFs that 206 only differ in UTRs for L4-100K, L4-33K, L4-pVIII, and E4orf6/7. In addition, we discovered 207 canonical ORF isoforms embedded within transcripts generated from non-canonical promoters, 208 The complexity of the adenovirus transcriptome has not been fully explored using modern high-249 throughput technologies. Here we integrate short-read cDNA sequencing and long-read direct 250 RNA sequencing to re-annotate both the Ad5 DNA genome and RNA transcriptome. Using high 251 quality and high depth short-read sequencing, we were able to detect SNPs within the transcribed 252 regions of the genome approaching 100% penetrance, indicating that these sites were likely 253 present in the genome and not due to RNA editing or modifications. We recapitulated the known 254 TSS and CPAS sites throughout the Ad5 genome, and annotated novel splicing events within the 255 viral transcriptome. Of these 20 novel RNAs, 13 are likely to encode for altered ORFs including 256 multiple fusion transcripts that span transcriptional units thought previously to be distinct. Overall, 257