v0.2.0

What's Changed

The workflow has changed substantially since v0.1.0 and is now as follows:

• Use StringTie to assemble transcripts from aligned RNA-seq reads
  Filter assembled transcripts for minimum mean expression across samples of the same condition (Biorxiv, Swamy et al., 2020)
• Filter for novel last exons that extend a known exon (optionally checking that the 5'ends match) or contain a novel last splice junction (optionally checking that the novel 5'ss matches a known 5'ss)
• Filter assembled transcripts for last exons with nearby reference polyA site (e.g. PolyASite) or conserved polyA signal motif (e.g. 18 defined in Gruber et al., 2016)
• Merge novel last exon isoforms with reference annotation, quantify transcripts with Salmon
• Output count/TPM matrices via tximport for use with downstream differential transcript usage packages
• Perform differential usage analysis between two conditions using DEXSeq

Other updates:

• Extensive documentation update
• Added license
• Addition of simulated test data
• sge profile removed (it was never working in the first place)

The relevant PRs can be found here:

• Script to filter GTF file based on attributes by @SamBryce-Smith in #29
• v0.2.0 by @SamBryce-Smith in #39
  Full Changelog: v0.1.0...v0.2.0

v0.1.0

Initial implementation of PAPA pipeline, which does the following:

• Assemble transcripts with StringTie
• Filter predicted transcripts based on mean expression, reference matching splice junctions (expect final novel SJ) and 3'end proximity to reference polyA site or polyA signal motif
• Merge filtered novel reference transcripts, quantify with Salmon
• Sum transcript expression to shared polyA sites/last exon isoforms
• Generate count & TPM matrices of per-polyA isoform counts for use with downstream differential transcript usage packages