Analysis code and supporting data for GFRAL signal comparisons in hyperemesis gravidarum

Analysis code and author-generated supporting data for molecular comparisons across six hyperemesis gravidarum (HG) loci and component-specific GFRAL analyses. Version 1.0.7 integrates sample-count reconstruction and scalar-N sensitivity analyses (Data S20), complete Fenland SomaScan GFRAL analyses (Data S21), and current numerical inputs and rendering code for Figures 1–5 (Data S17). The archive retains the initial molecular screens, matched-variant model comparisons, 54 effect-prior comparisons, 144 CSF scale/prior comparisons, and complete supporting, non-supporting and uncertain results. Fenland rs60676898 yielded a positive single-instrument MR association (P = 0.00175838; two-test BH q = 0.00351675); the original instrument rs72988065 remained nonsignificant with Fenland exposure estimates (P = 0.109099). Both reuse the same HG outcome. Conditional secondary-component colocalization does not establish primary-component replication or validate the new MR instrument. Conventional and exploratory LD-mismatch results, nonconverged fits, and the palindrome-exclusion sensitivity are distinguished explicitly. No causal or therapeutic-target validation is claimed. One ZIP contains the code, derived results, source manifests and reproduction instructions. Manuscripts, publication images, formatted workbooks, individual-level genotypes, access credentials and restricted provider datasets are not included. Author-generated code is under the MIT License; derived data and documentation are under CC BY 4.0. Third-party source data remain subject to provider terms. Full upstream analyses require separately obtained provider inputs and preparation steps described in REPRODUCE.md.