Skip to content

ISOFOX: integrate tars support (AUS-418) - #845

Merged
charlesshale merged 74 commits into
masterfrom
AUS418-isofox-tars
Sep 24, 2026
Merged

charlesshale merged 74 commits into
masterfrom
AUS418-isofox-tars

Conversation

@shiv-hartwig

@shiv-hartwig shiv-hartwig commented Jul 13, 2026 •

Copy link
Copy Markdown
Contributor

Functional:

  • only works with Tars and Redux, not STAR BAMs
  • removed all enriched gene logic, and enriched gene perc from statistics file
  • alt-sj check splice junctions vs known to ensure novel is selected where multiple exist in a read
  • fusion soft-clips reads no longer have a max bases of 10
  • fusion supporting fragments include junction and discordant, not realigned
  • fusion reads are more permissive keeping trans-exon refs related to an exon
  • fusion hard filter if no supplementary has map qual >= 10 (not applied to known pairs and promiscous genes)
  • fusion check for position adjustment from junction base overlap between primary and supplementary from BWA

Bugs:

  • alt-SJs skip homology shift only if both ends of splice junction match the same transcript's exon boundaries
  • trim adapter sequences prior to testing exon boundaries
  • trim adapters based on matched alignment positions, not read ends (to handle inconsistent split reads in fragment)
  • fusion fragments from primary and supp ensure the same pairing is used

Technical:

  • no slice of duplicates, instead extract duplicate count from consensus read attribute

Config:

  • removed RNA excluded regions config, now handled by Tars and Redux

@shiv-hartwig
shiv-hartwig force-pushed the AUS418-isofox-tars branch 3 times, most recently from e55c46a to d7b8e42 Compare July 21, 2026 04:02
charlesshale and others added 4 commits July 21, 2026 14:40
- add missing check for consensus reads
- don't trust chimeric-flagged reads from STAR
- add soft-clip within-exon threshold for alt classification
@shiv-hartwig
shiv-hartwig force-pushed the AUS418-isofox-tars branch 2 times, most recently from 3676dc3 to dc6abba Compare August 26, 2026 00:28
charlesshale and others added 23 commits September 14, 2026 19:45
- fusion read group checks all reads received in correct / standard manner
- fusion fragments from primary+supp ensure the same pairing is used
… based on matched alignment positions, not read ends
@charlesshale
charlesshale merged commit 56bb971 into master Sep 24, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants