Epigenetic testing: what methylation can—and cannot—say about mitochondria
Environmental signals can change gene regulation without changing the DNA sequence. That does not mean a blood or saliva clock can identify every exposure, inspect another tissue or diagnose mitochondrial dysfunction.
Three buckets—not one marketing story
Nuclear methylation regulates genes such as PGC-1α and TFAM that build and run mitochondria. Exercise can remodel those marks in human muscle.
Whether methylation of the mitochondrial genome itself is abundant and functional. Circular-DNA and NUMT artifacts explain much of the reported signal.
No consumer methylation test is validated to report ATP production, respiration or mitochondrial function in an individual.
The environment can regulate nuclear mitochondrial genes
Barrès et al. (2009) found PGC-1α promoter hypermethylation in type 2 diabetic muscle, negatively correlated with PGC-1α mRNA and mtDNA content. TNF-α and free fatty acids acutely increased non-CpG methylation in human myotubes; silencing DNMT3B prevented it. In 2012, a single exercise bout reduced promoter methylation of PGC-1α, PPAR-δ, PDK4 and TFAM in human muscle biopsies. The mechanism is real and reversible—but the evidence comes largely from muscle and cell models, while consumer tests read blood or saliva. Exercise study ↗
mtDNA methylation may be signal, artifact—or both
Evidence for function includes mitochondrial DNMT1 (Shock 2011), non-CpG D-loop methylation (Bellizzi 2013) and the newer adenine mark 6mA installed by METTL4, which can alter TFAM binding and transcription. The countercase is strong: linearizing circular mtDNA before bisulfite conversion collapsed apparent methylation from 15–32% to below 1.5%; nanopore analysis found no CpG methylation after controlling technical bias and nuclear mitochondrial DNA segments. In 2025, cigarette smoke impaired basal, ATP-linked and maximal respiration in human airway cells with no meaningful mtDNA-methylation change. The honest synthesis is unresolved, with 6mA/METTL4 the more promising early lead. mtDNMT1 study ↗ Bisulfite-artifact study ↗ Nanopore study ↗
Clocks are proxies, not mitochondrial diagnostics
In 6,316 people, GrimAge, DunedinPACE and other clocks correlated weakly with mtDNA copy number—a cross-sectional measure of mitochondrial content, not function. No clock was trained on ATP, respiration or another mitochondrial endpoint. TruDiagnostic’s TruHealth report markets “Metabolic & Mitochondrial Function” scores including ATP synthesis and NAD+ metabolism. Those are methylation-derived surrogates, not measurements, and no peer-reviewed validation against ATP production, respirometry or any mitochondrial endpoint was found. That is an unvalidated marketing claim. TruHealth claim ↗
The consumer testing landscape
TruDiagnostic TruAge, Elysium Index, myDNAge and similar services are laboratory-developed tests; none is FDA-cleared. Epigenetic clocks are validated most strongly for population-level age and mortality associations, not individual clinical decisions. Identical-sample deviations of up to roughly nine years, proprietary algorithms and tissue mismatch limit interpretation. CLIA certification speaks to laboratory quality—not the clinical validity of a clock or a “mitochondrial function” score.
False reassurance, false alarm, unnecessary medical cascades, privacy exposure and supplement/retest loops are foreseeable. TruDiagnostic sells a roughly $998/year quarterly retest subscription; Elysium sells Basis supplements alongside Index; TALLY was co-founded by David Sinclair. Those conflicts do not invalidate all clock science, but they matter when scores become sales funnels.