NEWS

---

gpmapr 1.0.0 (2026-11-07)

First CRAN release. Requires version 1.1.0 of the GPMap API.

• An R client for the 'OpenGWAS' Genotype-Phenotype Map API (https://gpmap.opengwas.io): look up traits, genes, variants and regions with their colocalisation groups, rare-variant results and associations; search the map with search_gpmap(); fetch LD proxies, pleiotropy summaries and duplicate traits; upload, fetch and delete your own GWAS.
• Requests that hit the API rate limit (HTTP 429) are retried with exponential backoff.
• The API version is checked on the first request of a session rather than when the package loads, so library(gpmapr) needs no network access. If the installed gpmapr does not support the API version, the first request stops with instructions to update the package.
• API errors and connection failures now give an informative error message.
• New gpmap_available() reports whether the API can be reached.
• search_gpmap() for a variant now reports a per-row number of colocalising studies and rare results (previously every row showed the first row's counts), and gains an rsq column with each proxy's r-squared with the searched variant. Variant searches no longer download the full list of traits and genes.
• Breaking: GWAS upload results (get_gwas(), and trait() / traits() with a GUID) have the same tables and column names as other results.
  • One study_extractions table holds the upload's study extractions and the existing ones it colocalises with; upload_study_extractions is gone.
  • Rows of study_extractions and coloc_groups have a source column ("upload" or "existing") and a gwas_upload_id (the upload the row comes from; NA for existing rows). Upload and existing study extractions have separate ids, so identify one by source and its id together. The upload's own rows are source == "upload" & gwas_upload_id == result$trait$id.
  • Coloc groups use study_extraction_id for both sides, and coloc pairs use study_extraction_a_id / study_extraction_b_id with source_a / source_b, replacing the upload_* / existing_* id columns.
  • Upload rows are named after their upload (trait_name), have data_type "Phenotype" and no trait_id or study_id.
• New compare_upload() lines up a GWAS upload with another trait (or a compared upload): each one's effect in every coloc group they share, and the coloc pairs between them.
• Coloc pairs from gene(), genes(), variant() and get_gwas() include the trait name, data type, tissue and gene of both sides (trait_name_a, trait_name_b, ...), so they no longer need joining to study_extractions.
• Breaking: trait is now pipeline_id (the GPMap pipeline's id for the trait, such as its OpenGWAS dataset id) and category is now variable_type (continuous or categorical), in all_traits(), trait(), traits() and GWAS uploads.
• Breaking: $associations is no longer returned. With include_associations = TRUE, each study's association at its variant (beta, se, p, eaf, imputed) is added to coloc_groups, rare_results and study_extractions. get_gwas() now adds them too, as trait(<guid>) already did, and so does traits() for GUIDs.
• Associations of variants a study did not report are placeholders in the GPMap (beta 0, se 1, p 1); they are now returned with beta, se, p and eaf set to NA.
• upload_gwas() takes variable_type in place of category, which still works but is deprecated.
• traits() no longer errors when given a GWAS upload GUID alongside trait ids.
• ld_proxies() and ld_matrix() gain an rsq column. ld_matrix() returns an empty table when none of the variants are a lead variant and its proxy.
• genes() and traits() drop empty elements, as gene() and trait() do, and genes() no longer returns tissues.
• region(), gene(), trait(), variant() and their plural forms no longer error with include_associations = TRUE when the API returns no associations.