Analysis code for "Reproducible cerebrospinal fluid proteomic signatures across early neuronal α-synuclein disease stages in PPMI

R analysis code accompanying Weber, Zaplatnikov et al., "Reproducible cerebrospinal fluid proteomic signatures across early neuronal α-synuclein disease stages in PPMI". The study examines which cerebrospinal fluid (CSF) protein alterations across prodromal and early manifest neuronal α-synuclein disease replicate across two independently profiled cohorts from the Parkinson's Progression Markers Initiative (PPMI), both measured on the Olink Explore HT proximity extension assay platform: a discovery cohort and a validation cohort. Participants comprise healthy controls, prodromal individuals with hyposmia or isolated REM sleep behaviour disorder, and drug-naïve sporadic Parkinson's disease, staged by the neuronal α-synuclein disease integrated staging system (NSD-ISS) using CSF α-synuclein seed amplification assay and DAT-SPECT. The repository contains R scripts covering: cohort construction and NSD-ISS staging; within-product bridge normalisation between the two projects; protein-wise differential abundance with limma, adjusted for age, sex, plate and sample-level median NPX, for clinical subgroup and NSD-ISS stage contrasts; cross-cohort replication and DerSimonian–Laird random-effects meta-analysis; KEGG overrepresentation and gene-set enrichment analysis with cross-cohort concordance testing; covariate-residualised partial correlations of protein abundance with DAT striatal binding ratio, MDS-UPDRS III and UPSIT; and a phenoconversion analysis combining longitudinal decline-rate models with elastic-net Cox stability selection under repeated nested cross-validation, including random-panel and outcome-permutation null controls. No data are included in this deposit. PPMI data are controlled access and are obtained from the PPMI LONI portal (https://www.ppmi-info.org/access-data-specimens/download-data, RRID:SCR_006431) under a Data Use Agreement; DaTscan and α-synuclein SAA results for prodromal participants require separate approval by the PPMI Data Access Committee. The scripts read from a local raw-data directory configured by the user. Developed in R 4.5.1. Key dependencies: limma, OlinkAnalyze, clusterProfiler, survival, glmnet, metafor, openxlsx.