Skip to content

Releases: lcpilling/ukbrapR

ukbrapR v0.3.14

Choose a tag to compare

@lcpilling lcpilling released this 02 Mar 14:28

Bug fixes

  • get_df() now removes any duplicated IDs if using the death data to get the censoring date (this unfortunately resulted in some extra rows added during a later left join).

ukbrapR v0.3.13

Choose a tag to compare

@lcpilling lcpilling released this 21 Feb 23:28

Bug fixes

  • Fixed issue #39 for create_pgs() to correctly load the new PGS from plink2 (thanks to @aepacker for highlighting)

Changes

ukbrapR v0.3.12

Choose a tag to compare

@lcpilling lcpilling released this 16 Feb 11:28

Bug fixes

  • get_df() fixed where no results are returned for a condition - previously returned an error, now just returns a binary variable with all 0's for the condition

ukbrapR v0.3.11

Choose a tag to compare

@lcpilling lcpilling released this 10 Feb 08:58
98ff2ef
  • Use plink2 when creating polygenic scores. This handles effect/other allele differences more robustly
    • Returns the "scoreavgs" which is the total sum of the effect weights divided by the number of alleles
  • Export "GP registrations" table when submitting table-exporter commands

ukbrapR v0.3.10

Choose a tag to compare

@lcpilling lcpilling released this 27 Jan 21:19
6f9bb7e

New features

  • For genetic functions using the imputed bgens it is now possible to extract based on position (build 37) not RSID. For example:
    • In extract_variants() or create_pgs() when source="imputed" new option use_imp_pos can be set to TRUE (default is false)
    • This correspondingly sets use_pos to TRUE when calling make_imputed_bgen() internally (to extract based on position not RSID)
  • New function get_imputed_variant_info(): for a given set of genomic coordinates (position in build 37) get the UK Biobank imputed genotype variant IDs from the MFI files (also returns alleles, MAF and INFO).

Thanks to Harry Green @hdg204 for the idea and bug testing.

ukbrapR v0.3.9

Choose a tag to compare

@lcpilling lcpilling released this 04 Dec 13:08
c607e14

Changes

  • Fix issue with dx paths when downloading exported data in /ukbrapr_data. Previous method worked fine in RStudio but not Cloud Workstation, and highlighted a dependency on relative paths for $ dx download internal commands.
  • Add {tidyr} package to requirements. Has always been a requirement, but now it is explicit!

ukbrapR v0.3.8

Choose a tag to compare

@lcpilling lcpilling released this 02 Dec 16:14
34b9af1

Changes

  • Implement suggestion by @nsandau (issue #34) to use --brief --yes flags when exporting tables with export_tables() (if running from terminal user was prompted for each job).

ukbrapR v0.3.7

Choose a tag to compare

@lcpilling lcpilling released this 20 Jun 08:36
bf4baf1

Changes

  • make_XXX_bed() functions now give a more accurate ETA. Previously only considered the number of CHRs to search. Now considers the number of variants also. -- updated to be more accurate (sometimes overestimated substantially)

Bug fixes

  • Fix issue #31 where fields_to_phenos() incorrectly arrayed some "multiple choice" questions. Thanks to Jana Sönksen for highlighting

ukbrapR v0.3.6

Choose a tag to compare

@lcpilling lcpilling released this 29 May 21:03
bade2c0
  • make_dragen_bed() returns a more informative error if DRAGEN BGENs not found (suggests user is in a Project not updated since April 2025).
  • make_XXX_bed() functions now give a more accurate ETA. Previously only considered the number of CHRs to search. Now considers the number of variants also.
  • Removed indy/snow paths (i.e., my local servers) because we are now fully RAP working.
  • Removed Windows functionality. The RAP is always UNIX.

ukbrapR v0.3.5

Choose a tag to compare

@lcpilling lcpilling released this 16 Apr 13:53
970a2c5

New features

  • create_pgs() can now use a local BED file of already-extracted variants. Before, it assumed the user wanted to extract the variants prior to creating the PGS.