Releases: bcgsc/ntRoot
Releases · bcgsc/ntRoot
ntRoot v1.1.6
- Release v1.1.6
- Added changes to be compatible with ntEdit v2.1.1+, where the reads prefix (
--reads
) can now be a relative or absolute path to the reads directory- Backwards compatible - if no path specified, it will search the current working directory for input files as before
ntRoot v1.1.5
- Release v1.1.5
- Fix for
--custom_vcf
mode- Allow variant (non-header) lines in the input VCF to have internal
#
characters
- Allow variant (non-header) lines in the input VCF to have internal
ntRoot v1.1.4
- Release v1.1.4
- Refactoring single-character parameters of
ntroot_run_pipeline.smk
- Newer snakemake versions fail to properly parse single-character config keys
- Use multi-character parameter strings to ensure compatibility
ntRoot v1.1.3
- Release v1.1.3
- Added flexibility for parsing input VCFs when using
--custom_vcf
- Accepts 8-10 columns, auto-detects for parsing bedtools intersect file
- Previously, required 10 columns (as were present in all VCFs tested), but it is valid to have as few as 8 (as there are 8 required columns as per VCF specifications)
- If there are more than 10 columns, this would indicate a multi-sample VCF, which doesn't make a lot of sense for using ntRoot to evaluate an individual's ancestry - in this case, an informative error message is printed
- Accepts 8-10 columns, auto-detects for parsing bedtools intersect file
ntRoot v1.1.2
- Release v1.1.2
- Added an "Unknown" category to ancestry predictions
- For cases where a given genomic tile has ambiguous ancestry due to multiple populations having an identical highest score
ntRoot v1.1.1
- Release v1.1.1
- Fixes to workflow so
--lai
mode is compatible with gzipped reference fasta files - Mention name of tile resolution output file in ntRootAncestryPredictor.pl in
--lai
mode
ntRoot v1.1.0
- Release v1.1.0
- Added feature to allow for inferring ancestry from a provided VCF
- When
--custom_vcf
is specified, ntRoot will use this VCF for inferring ancestry instead of running ntEdit - When using an ntEdit VCF file with a new
-l
variant call set, use--strip_info
to ensure that there are no collisions between existing cross-referenced allele assignments in the ntEdit VCF and new cross-referenced alleles
- When
- Updated ancestry inference script to allow for comma-separated allele frequencies when alt alleles exist
- Documentation updates
ntRoot v1.0.1
- Release v1.0.1
- Make ntEdit2 a dependency instead of using it as a submodule
- The meson build system was removed due to there no longer being any compiled components of ntRoot
- Documentation updates
ntRoot v1.0.0
Initial release of ntRoot: Ancestry inference from genomic data