LRAA Release v0.12.0
·
320 commits
to main
since this release
v0.12.0, November 14, 2025
- caching initially constructed splice graphs for reuse in downstream quants to avoid reconstruction effort.
- all splice graphs for all steps leverage the normalized read coverage (not just isoform ID)
- added --num_parallel_contigs to control contig-level concurrency (default 4 workers); --num_threads_per_worker (legacy --CPU alias) specifies cores allocated per contig/strand worker (default 1 core)
- default --oversimplify now targets mitochondrial contigs ("chrM,M")
- wdls optimized for splitting data per chromosome
- scatter runs per chromosome only (strand-level splitting removed) with subjobs shuffled for improved load balancing
- including ref-annot gene symbol integration into single cell pipeline