v0.5.3

Changelog

• Add more sources for the harmonized scoring file validator by @ens-lgil in #74

Full Changelog: v0.5.2...v0.5.3

v0.5.2

Changelog

Bug fixes

• Fix parsing optional is_dominant and is_recessive fields in some PGS Catalog scoring files (e.g. PGS000802)
• Optimize speed and reliability of ancestry_analysis
• Fix handling rare cases when adjusting a PGS with 0 variance (skip)

v0.5.1

Changelog

Improvements

• Make aggregate_scores output consistent with ancestry_analysis output

Bug fixes

• Fix aggregate_scores not respecting --outdir
• Bump dependency versions with high security vulnerabilities

v0.5.0

Changelog

• Refactor combine_scorefiles
• Deprecate samplesheet module
• Add custom exceptions

v0.4.3

Changelog

• Fix retrying downloads when the EBI servers are sleepy on a Monday morning
• Give a more helpful error message when there's no valid matches in match_combine
• Fix numeric sample identifiers breaking ancestry analysis
• Check chr prefix in samplesheets

v0.4.2

Changelog

• Fix relative path discovery behaviour.
  • Before: relative to grandparent of input file (work directory)
  • Now: relative to input file, resolving symlink if present
• Check for reserved tokens in sampleset names
• Update dependencies to newer version

v0.4.1

Changelog

• fix checksum warning when downloading scoring files #53
• fix custom user-agent when querying PGS Catalog API and downloading scoring files #54

v0.4.0

Changelog

New features

• ancestry_analysis: use genetic PCA loadings to compare samples to population reference panels, and report PGS according to where they sit in empirical distributions (percentile/Z) and/or PGS adjusted for continuous axes of genetic ancestry (regressing PGS~PCs).
• relabel_ids: a program to make a variant information file IDs match another variant information file (e.g. a reference panel)
• check_samplesheet: a program for validating pgsc_calc samplesheets

Improvements

• download_scorefiles refactored to prefer HTTPS and attempt FTP fallback (to minimise FTP busy errors)

v0.3.1

Changelog

Bug fixes

• Fix reading effect weight column of some PGS Catalog scoring files as a string instead of a float (e.g. PGS002619, PGS000908, PGS002704, PGS002678) PGScatalog/pgsc_calc#79

v0.3.0

Changelog

• Handle large traits in the PGS Catalog more efficiently (e.g. cancer, autoimmune diseases)
  • Split - apply - combine approach: added combine_matches for use with (match_variants --only_match)
  • More lazy operations
• Improve correctness of labelling: now all matches are labelled simultaneously
• Get scoring file information (e.g. publication, trait) from header data and store in a separate log file