A smoothing method for DNA methylome analysis to enhance epigenomic signature detection in epigenome-wide association studies
Bibliographic record
Abstract
Epigenome‐wide association studies (EWAS) are instrumental for mapping DNA methylation changes in human traits and diseases but often suffer from low statistical power and false positives, especially in small cohorts. We developed an EWAS smoothing method that exploits co‐methylation of adjacent CpG probes within CpG islands via a sliding‐window average and generalized it using Savitzky-Golay filtering. We applied the smoothing approach—with window widths of 1–3 CpGs and, for generalization, Savitzky-Golay filters of varying polynomial orders and window sizes—across five distinct EWAS settings. Performance was quantified by signal‐to‐noise ratio (SNR), noise‐variance reduction, variance ratio (VR), Bayes factors, and sample‐size sensitivity. In the MMACHC epimutation dataset, a 5‐CpG window (width, w = 2) increased SNR by 90 %, reduced noise variance by 80 %, and elevated VR by 176 % at the target CpG island, with no genome‐wide false positives. For MLH1 , smoothing preserved the top association and suppressed background signals. In the aging EWAS, a “Polyepigenetic CpG aging score” was derived following smoothing. This score correlated strongly with chronological age in the discovery cohort (Spearman’s ρ = 0.89; P = 3.0 × 10 −219 ) and was independently validated in a separate dataset, significantly distinguishing newborns from nonagenarians ( P = 3.4 × 10 −8 ). Savitzky-Golay filtering of order 0 with a 5‐CpG window yielded optimal SNR across bootstrap iterations, supporting this configuration as a robust choice for methylation array smoothing. As an extension of the Savitzky-Golay-based smoothing framework, reanalysis of a liver cancer dataset identified five top loci surpassing a smoothed P -value threshold of 1 × 10 −8 . Among these, MIR10A within the HOXB3 locus was the only previously reported functionally relevant site. In conclusion, the smoothing method improves EWAS performance by enhancing SNR, enabling detection of meaningful associations even in small cohorts, and offers a valuable tool for reanalyzing existing Infinium methylation array datasets to uncover previously undetected epigenomic signatures.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.016 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.005 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".