Gallardo, C. M.; Albert, J. L.; Qazi, A. A.; Ventura, R. L.; Deshmukh, S.; Beliakova-Bethell, N.; Torbett, B. E.

Alternative splicing (AS) greatly expands the repertoire of proteins encoded by the human genome. Viruses have been shown to hijack AS to sustain replication or lead to latency. In HIV infection, the virus integrates into the host genome, becoming a transcriptional unit that directly engages in AS to regulate its gene expression. Sequencing advances have enabled insights into HIV gene expression dynamics during productive replication. However, viral isoform dynamics during latency remain largely uncharacterized due to the low abundance of both spliced viral transcripts and their originating cellular subsets, making their accurate detection and quantification challenging. To address current gaps, we developed MrHAMER2, a high-accuracy long-read RNA sequencing method that leverages dual Unique Molecular Identifier (UMI) tagging of cDNA to accurately capture and quantify full-length isoforms with high dynamic range and 99.968% single-nucleotide accuracy. We then use MrHAMER2 to decode the spliced HIV transcriptome in a primary cell model of latency and show substantial changes in viral isoforms bearing intron retentions accompanied by changes in their potential to generate translatable protein.