Convergence Without Coherence: A Pre-Registered Test of Distributed Transdiagnostic Epigenetic Architecture in 181 Adversity-Related EWAS

A pre-registered test of distributed transdiagnostic epigenetic convergence in 181 adversity-related EWAS, using public summary-level data from the EWAS Catalog (snapshot 2026-04-09). The toolkit identifies CpG loci with directionally convergent effects across independent studies under pre-registered thresholds (concordance ≥75%, |median Δβ| ≥0.01, ≥3 studies), tests the resulting gene set for functional enrichment in Gene Ontology Biological Process and Reactome, and compares it against 1,000 size-matched random gene sets sampled from the EWAS Catalog background. A tissue-stratified robustness analysis restricted to adult blood-tissue studies is also performed. The pre-registered hypothesis of a distributed, functionally coherent epigenetic architecture detectable in adversity-related EWAS was refuted under the study's pre-registered decision criteria. Although 828 directionally convergent CpG loci were identified across the adversity corpus, the resulting gene set produced fewer significantly enriched biological terms than size-matched random gene sets and showed no evidence of functional coherence beyond random expectation. The most represented contributors (PRDM8, PPT2, PRRT1, AHRR, GFI1, and CYP1A1) include canonical EWAS signatures of tobacco smoking. The adult-blood robustness analysis further supports smoking as a plausible major confounder contributing to apparent cross-study convergence. The repository contains the complete reproducible research workflow, including the Python analysis pipeline (convergence_analysis.py), the LaTeX manuscript (main.tex), the primary figure (figure1.pdf), the pre-registration document (preregistration.md), intermediate outputs, and licensing information. The software is released under the MIT License, while the manuscript and figures are distributed under CC BY 4.0. This work is intended as a reproducible methodological evaluation of convergence claims in adversity-related EWAS and provides all materials necessary for independent verification, reanalysis, and extension. Independent research conducted without institutional affiliation or external funding. Author: Verónica Zumpano Blumenfeld ORCID: 0009-0006-2030-1849 0009 https://github.com/Vlattice