v0.4.3
Fixes sorting error in workflow that caused ngsF-HMM to fail when underscores were in contig names and enables including | in contig names.
!!! Updating to this version from a previous version will potentially trigger reruns of certain rules
Results from previously successful runs will not change during these reruns, so they can be ignored if you have not previously experienced these errors. More details:
To fix some bugs in how contigs with special characters were handled, two rules were updated: ref_chunking, which creates lists of groups of contigs to parallelize analyses over, and prune_chunk_beagle, which subsets beagle files after sites have been pruned to include only the pruned sites. Since ref_chunking is an early rule in the workflow, rerunning it would trigger reruns of most steps after mapping, so consider whether you can ignore the reruns:
- If your contig names all contain only alphanumeric characters or
., you likely do not need to rerun anything, as bugs did not impact contigs with only these characters. - If your contig names contain
_, you can safely ignore reruns ofref_chunking, but you should rerunprune_chunk_beagleif you want to run ngsF-HMM. The main fix here is maintaining site order after pruning for ngsF-HMM inputs containing special characters. While the pruned sites are also out of order for PCA and Admixture analyses, order should not influence the results here, so you do not need to rerun if you only target these outputs.
If your contig names contain |, your run would have previously failed at ref_chunking. This fix now allows this character.
What's Changed
- fix: Beagle files retain site order when pruned by @zjnolen in #67
- fix: Prevent reference chunking failures when contigs have bash characters (e.g. pipes) by @zjnolen in #69
Full Changelog: v0.4.2...v0.4.3