GIA: Germline-Informed Aging with AlphaGenome Finds Genetically Regulated CpGs
q-bio.QM · q-bio.GN
Quick verdict
Reports 2.6-fold meQTL enrichment of clock CpGs; matched analysis is sound, but the FHL2 highlight is exploratory and no effect on clock estimates is tested.
Summary
The paper asks whether CpGs used by published epigenetic clocks are preferentially under germline genetic control. GIA compiles 3,837 unique CpGs from 13 clocks via Biolearn, queries GoDMC for independent blood meQTLs, and compares hit rates against 1:1 matched unused Illumina 450k probes (same chromosome and CpG-island class, fixed seed). 62% of clock CpGs had a blood meQTL versus 39.6% of matched probes (OR 2.57, Fisher P=3.8e-90); the enrichment held per-clock (11/13 FDR-significant) and was extreme among the 16 CpGs used by 5+ clocks (15/16 vs 4/16, OR 45). Clock lead SNPs also showed larger |beta| and more independent meQTLs but relatively fewer cis associations. Fifteen recurrent-site lead SNPs scored with AlphaGenome on blood RNA-seq/DNase/ATAC/TF-ChIP tracks were indistinguishable from 15 matched meQTL SNPs after FDR (all q>=0.27); ranking by raw chromatin score instead highlighted rs10190186, a cis FHL2 variant predicted to raise accessibility and FHL2 expression and associated with higher methylation at two nearby clock CpGs. Key limitations: the genetics-vs-aging variance decomposition motivating the work is never actually performed; controls are matched only on chromosome and island class; the n=15 AlphaGenome comparison is underpowered and the FHL2 call is an uncorrected post-hoc ranking; no repository URL is given.
Review complete
Assessment at a glance
5
risk flags identified
● high 0
● medium 3
● low 2
5 strengths · 5 weaknesses · 5 author questions
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Clock CpGs are enriched about 2.6-fold for independent blood meQTLs relative to chromosome- and island-class-matched unused 450k probes (62.7% vs 39.6%; Fisher P=3.8e-90).
- The enrichment holds within every clock family (11/13 clocks FDR-significant) and is strongest among the 16 CpGs used by five or more clocks (15/16 vs 4/16; OR 45.0).
- Lead SNPs at recurrent clock sites show no class-wide AlphaGenome regulatory enrichment versus matched blood meQTL SNPs after FDR correction (all q>=0.27).
- rs10190186, a cis variant at FHL2, is the top-ranked recurrent-site lead SNP by predicted chromatin effect, and its G allele raises methylation at two nearby clock CpGs in the direction of their reported age-related hypermethylation.
Strengths
Well-specified matched-null design
Unused 450k probes are paired 1:1 on chromosome and CpG-island class with a fixed random seed (Section 2.4), directly addressing the array-composition confound that naive enrichments ignore; fallback rules for sparse strata are stated.
Circularity-aware clock selection
Section 2.2 excludes Ying CausAge/DamAge/AdaptAge because they 'select meQTLs by design', and documents all other exclusions (duplicating clocks, nested GP-age models, 1-3 CpG clocks), making the 13-clock panel defensible rather than cherry-picked.
Robustness across clock families
Honest negative result at the SNP layer
Appropriate statistics with effect sizes
Weaknesses
The motivating question is never tested
Limited control matching
Underpowered, post-hoc SNP layer
Enrichment conflated with influence
Reproducibility pointer incomplete
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
- All 5 strengths & 5 weaknesses, analyzed in detail
- 5 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
Do epigenetic clocks measure aging or inherited genetics?
Both can contribute. This study found that 62.7% of the CpG sites used by 13 published epigenetic clocks have genetically regulated methylation in blood (meQTLs), about 2.6 times more often than matched unused probes from the same array. That means some differences between people's clock readings could reflect inherited baseline methylation rather than aging itself — although the paper did not measure how much of the clock output genetics actually explains.
What is rs10190186 and why does it matter for aging clocks?
rs10190186 is a single-letter DNA variant (A>G) inside the FHL2 gene. It is the strongest blood meQTL for two nearby CpGs each reused by four to five epigenetic clocks, and AlphaGenome predicts it increases blood chromatin accessibility and FHL2 gene expression. Its G allele is associated with higher methylation at both CpGs, matching their known age-related hypermethylation — making it a candidate inherited modifier of clock inputs, though this remains unvalidated.
What data and tools did the GIA study use?
Only public resources: CpG lists from Biolearn for 13 published epigenetic clocks (including Horvath 2013, Hannum, PhenoAge, GrimAge V1, DunedinPACE, and DNAmTL), the GoDMC catalog of blood methylation quantitative trait loci, Ensembl for coordinate lifting, and AlphaGenome for variant-effect predictions on blood regulatory tracks.
Did the study find that clock-associated SNPs have unusual regulatory effects?
No. Lead SNPs at recurrently used clock CpGs were statistically indistinguishable from other blood meQTL SNPs across ten AlphaGenome blood summaries after FDR correction (all q>=0.27). The FHL2 variant emerged only from an exploratory ranking by raw predicted chromatin effect and awaits fine-mapping and functional validation.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Figures 1-4 were available only as embedded captions and text descriptions; the visual content could not be independently verified.
- The 'project repository' is referenced without a URL in the extracted text, so code/data availability could not be confirmed.
- Supplementary material beyond Supplementary Table S1 was not included in the extraction.
- AlphaGenome and the AlphaGenome Atlas (references 6-7, dated 2026) could not be cross-checked; the SNP-layer results rest entirely on those predictions.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.