Fig. 2

3′ ends of BLV antisense transcripts. Ideogram of the BLV provirus, with 3′RACE reads for L267 mapping to the BLV genome shown below. Potential 3′ ends of the BLV antisense transcripts are shown. a Zoomed in IGV screen shot showing a large cluster of reads with poly A tails (shown as T in IGV) adjacent to the end of miR-B5-3p. No signal position predominates, although the majority were located downstream (in the context of the AS transcript) of a canonical AAUAAA polyadenylation signal sequence (PAS). AS1 transcripts ending at this position were given the designation AS1-S. b In five of the libraries subjected to 3′RACE a small number of poly A reads were observed at position ~5237. This is located 20 bp downstream (in the context of the AS transcript) of a second canonical AAUAAA (PAS). Further, 3′RACE carried out on YR2 with the forward primer adjacent to this potential poly A site rather than the in AS1 exon 1 also showed polyadenylated products matching this position. c For AS2, only L267 showed any evidence of a poly A tail, however no canonical PAS was observed and no consistent poly A site was observed. In the remaining samples (also observed in many L267 reads) the AS2 transcript underwent splicing with the host genome. Shown are the soft clipped bases (the bases in the read that did not map to BLV) for YR2 mapped to the ovine genome with BLAT (http://genome.ucsc.edu/). These bases correspond to an exon of the ELF2 transcript ~5 kb upstream of the YR2 provirus integrated site. d Ideogram of the BLV provirus, with 3′RACE products from YR2 sequenced on a MinION instrument (Oxford Nanopore Technologies) and mapping to the BLV genome. Location of antisense transcripts are also shown. As with Illumina sequencing the majority of the reads take the form of the shorter AS1-S, some reads do extend to the postulated end of AS1-L at position ~5237. A number of long reads end before this position. b Mapping of the soft clipped reads with BLAT (http://genome.ucsc.edu/) reveal that many of these transcripts are undergoing splicing with the ELF2 transcript. The large number of polymorphisms observed in the mapped reads are due to the high error rate observed in Nanopore reads