DNA Methylation Entropy as a Marker of Epigenetic Rejuvenation in Mouse Embryogenesis
- Publicada
- Servidor
- Preprints.org
- DOI
- 10.20944/preprints202606.2135.v1
Background/Objectives: DNA methylation dynamics are closely linked to ageing. While epigenetic clocks predict biological age from average methylation levels, they may miss structural rearrangements. The "epigenetic rejuvenation" hypothesis posits that embryonic biological age decreases toward a minimum ("ground zero") at gastrulation. This study tested whether DNA methylation entropy decreases from stage E4.5 to E6.5. Methods: Publicly available single-cell nucleosome, methylation and transcription sequencing (scNMT-seq) data (GSE121690; 364 mouse cells) were analysed. Five entropy measures were calculated: Shannon, Renyi, Tsallis, Lempel-Ziv (LZ) complexity, and local gradient entropy. Persistent entropy (PE) and topological analysis (Rips complex, persistence diagrams H₀ and H₁) were computed. A support vector machine (SVM) classifier was trained to distinguish developmental stages. Results: Mean Shannon entropy decreased from 0.841 (E4.5) to 0.805 (E6.5; p < 0.01). All five entropy measures and PE decreased significantly. LZ complexity showed the largest reduction (−28.4% binary, −11.4% ternary; p < 0.001). Ternary LZ correlated with Shannon entropy (r = 0.71). PE decreased from 15.91 to 14.89 (p = 0.01). Topological analysis revealed reorganisation: E6.5 showed stable H₁-cycles (9 intervals), absent at E4.5 (2 intervals), reflecting lineage-specific diversification. Regional entropy and disorder decreased (RE: −25.5%, RD: −27.4%, p < 10⁻¹³). The SVM achieved 93.4% accuracy (AUC = 0.981). Conclusions: The decrease in DNA methylation entropy from E4.5 to E6.5 supports epigenetic rejuvenation as an approach to "ground zero." Topological analysis revealed a qualitative reorganisation: E6.5 showed stable H₁-cycles (9 intervals) that were largely absent at E4.5 (2 intervals). Comprehensive characterisation with balanced subsampling revealed a significant decrease in max persistence (0.446 → 0.218, p = 0.03) and total persistence (2.170 → 1.070, p < 0.001), indicating a transition from a single dominant cycle to multiple smaller, evenly distributed cycles.