You can get full transcriptome annotation for the NL4-3 genomes: https://ccb.jhu.edu/HIV_Atlas/11676/AF324493.2 . Note the accession ID though - if you have a different one - either 1) search the HIV Atlas for your accession, 2) use the AF324493.2 instead (if you download from HVI Atlas it will provide you with both genome fasta and transcriptome annotation) or 3) use the Vira method to transfer HXB2 annotation (or AF324493.2 annotation) from the HIV Atlas onto your target genome assembly.

We have recently completed reference-grade annotation for several thousand HIV-1 genomes including the HXB2 (K03455.1), 89.6 and NL4-3 reference genomes: https://ccb.jhu.edu/HIV_Atlas/. Each annotation features full set of US, PS and FS messages along with protein assignment and major donor and acceptor sites. The annotations are provided in GTF and GFF formats, you can browse them directly on the web interface via integrated JBrowse2 and use them with any transcriptomic utilities you would use for human or other genomes (assembly, quantification, gene/tx expression, etc).

This project started from my personal need to improve spliced alignment with HISAT2/STAR and minimap around splice sites and to be able to compute transcript and junction expression effectively. Eventually, this took me down the rabbit hole of creating several methods for annotation transfer in HIV and annotation of thousands of LANL complete genome assemblies. You can also read more about the resource in our preprint

Hi Xiao,

Were you able to make progress on this? If so, could you share what you did/found?