Circulating serum proteins across the liver–heart axis in metabolic dysfunction-associated steatotic liver disease: a proteome-wide two-sample Mendelian randomization and colocalization study

Analysis code for a two-sample Mendelian randomization and colocalization study of the serum proteome in metabolic dysfunction-associated steatotic liver disease (MASLD). The pipeline screens 3,875 autosomal aptamers from the AGES-Reykjavik SomaScan v4 serum release against MASLD, coronary artery disease, heart failure, atrial fibrillation and any stroke, using cis-pQTL instruments only, and then tests whether each surviving protein-outcome association shares a causal variant with the outcome signal. Of 3,860 protein-outcome pairs, 225 in 183 proteins pass false-discovery correction and 66 are carried into colocalization. The deposit contains 22 numbered pipeline stages, seven shared modules, the derived inputs (cis windows, protein-to-gene map, instrument set, outcome extracts) that let the screen, the sensitivity analyses, the figures and the tables be rerun without re-downloading any summary statistics, and the reference outputs behind every table and figure of the article. Raw third-party summary statistics are not redistributed; the fetch stage re-downloads them. Requirements: Python 3.11 or newer with numpy, scipy, pandas and matplotlib; curl and a POSIX shell for the download stages. All estimators are implemented in the deposit and checked against closed-form answers, so results do not depend on the version of any MR package. Code is MIT licensed; derived tables are CC BY 4.0. Ethics: not applicable (publicly available summary statistics only).