AfricaRice Genebank germplasm distribution register 1995–2024 linked to the active collection: data and analysis code

Data and code for the article Which conserved rice diversity is requested? Thirty years of germplasm distribution from the AfricaRice Genebank linked to its full collection (Food and Energy Security, manuscript 2113995, revision R1). Contents: (1) the GRIN-Global Community Edition distribution register of the AfricaRice Genebank (128,331 rows, recipient person names removed); (2) the cleaned analytical dataset of 125,888 shipped accession records, 1995–2024, with harmonised recipients, corrected countries, recipient classes and shipment identifiers; (3) the collection frame of the 20,391 accessions of the active collection (Genesys, September 2026) with covariates (species, biological status, origin, acquisition year, Passport Data Completeness Index snapshots, mini-core, genotyping and characterization status, availability) and demand outcomes; (4) the institution authority table, identifier reconciliation and cleaning log; and (5) the full Python analysis pipeline (descriptive statistics, concentration measures, hurdle models with logistic and zero-truncated negative-binomial parts, sensitivity analyses and figures). See README.md for definitions, a data dictionary and reproduction steps. Notes: the register's intended-use field is not a request-level declaration and its MLS flag reflects current accession status rather than the transfer instrument; both are documented but not analysed as recorded. Version 1.1 adds the shipment-level case–control analysis (p9_casecontrol.py: logistic regression with shipment fixed effects and recipient-clustered errors; Bayesian mixed model with crossed shipment and recipient random effects) and the pedigree linkage of conserved accessions to improved lines recorded in the collection (p10_pedigree.py). Data files are unchanged from version 1.0.