HoldsUp

Morgan Levine

Altos Labs / Yale University School of Medicine

Morgan Levine, PhD, is an assistant professor of pathology at the Yale University School of Medicine and a Founding Principal Investigator at Altos Labs. Her research centers on the science of biological aging, epigenetics, and the development of aging metrics such as the PhenoAge clock. Her published work investigates epigenetic age acceleration, blood methylation tests across physiological systems, mitochondrial function, and the effects of lifestyle interventions on biological age.

37 claims checked on air: 1 contradicted 1 overstated 32 supported 3 unverified

What they said on air - citing their own research

6 citing their own research

0:10:42supportedhightheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

The PhenoAge epigenetic clock, published in 2018, was trained on clinical lab test markers combined into a mortality-predictive measure rather than directly on chronological age.

"So what the second-generation clocks did—the one that we published in 2018 was the first example—is we said, "Oh, can we come up with a better thing to try and kind of tune these measures to?" So in that case, we used kind of normal lab tests that we combined into a measure that was predictive of mortality, and then we trained a predictor of those lab tests." (said at 0:10:42)

The 2018 paper introducing DNAm PhenoAge by Levine et al. used a two-step process where clinical chemistry lab tests and chronological age were first combined into a composite clinical measure of phenotypic age calibrated to mortality risk, and then DNA methylation data were trained to predict this phenotypic measure rather than chronological age alone.

0:22:26supportedmoderatetheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

On average, females exhibit lower epigenetic age than chronological age-matched males.

"So on average, not across the board, but if you look at the distributions, females on average will have slower or lower epigenetic age than same chronological age males." (said at 0:22:26)

Multiple large-scale observational studies using various DNA methylation clocks (such as the Horvath, Hannum, and GrimAge clocks) demonstrate that females on average exhibit lower epigenetic age and slower epigenetic age acceleration compared to chronological age-matched males across multiple tissue types (including blood, saliva, and brain tissue).

0:23:00supportedmoderatetheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

Natural and surgical menopause are associated with accelerated epigenetic aging.

"So we looked at women who had undergone menopause and how long since they'd undergone menopause, and it seems to be that menopause is actually an epigenetic aging accelerated event. So before menopause, women are doing pretty well, and then when they go through menopause, it seems to accelerate their epigenetic age. And we didn't have the kind of data you would want where we'd have the same women pre- and post-, but we can even look at surgical menopause, and that seems to also show this kind of accelerated epigenetic aging manifestation." (said at 0:23:00)

Large multi-cohort observational analyses and Mendelian randomization studies confirm that both natural menopause (earlier age at menopause and longer time post-menopause) and surgical menopause (bilateral oophorectomy or hysterectomy) are significantly associated with accelerated epigenetic aging measured by DNA methylation clocks.

1:04:39supportedhightheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

Running the same split blood sample twice on original epigenetic clocks can produce differences of up to eight years in estimated epigenetic age.

"we've taken blood samples, you can split them, like the same sample run it twice, and you can get upwards of eight years difference in your epigenetic age using traditional clocks." (said at 1:04:39)

Published technical evaluations of traditional epigenetic clocks show that technical variation between replicate split samples from the same biological source can produce discrepancies of up to 8 to 9 years in estimated epigenetic age. Higgins-Chen and colleagues (2022) systematically evaluated six prominent original epigenetic clocks and demonstrated that technical noise causes replicate deviations of up to 9 years, leading to the development of principal-component-based clock variants to resolve this reliability issue.

1:05:09overstatedmoderatetheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

A statistical method that removes technical noise reduces the test-retest variation of split samples on epigenetic clocks to a maximum difference of about one year.

"we actually developed a statistical method that completely removes all this technical noise. And I won't go into the math for people on the podcast, but basically, we can get this down to: you can split the sample, and now you're getting only about one year difference at max." (said at 1:05:09)

A 2022 study by Higgins-Chen, Levine, and colleagues introduced principal-component-based (PC) epigenetic clocks to address technical noise in DNA methylation assays. In their validation across six prominent epigenetic clocks, the PC approach significantly reduced technical noise between split-sample replicates (from discrepancies of up to 9 years down to agreement for most replicates within 1.5 years). However, the claim that the method "completely removes all" technical noise and restricts differences to "at max" about one year overstates the findings, as residual variation remains and 1.5 years reflected the range for most, but not strictly all, replicates.

1:09:43supportedmoderatetheir own paperMorgan Levine, PhD, on PhenoAge and the Epigenetics of Age A

Re-analysis of Kara Fitzgerald's intervention dataset using statistical noise-removal methods showed that the observed reversal in epigenetic age was entirely attributable to technical noise.

"And we were actually able to go back in and show that the entire effect was noise. So actually, once you do the statistical method that removes the noise, there was actually no effect of the intervention." (said at 1:09:43)

Re-analysis of intervention trials (such as the diet and lifestyle intervention published by Fitzgerald et al.) using principal-component (PC) and reliability-adjusted epigenetic clock methods developed to remove technical noise (Higgins-Chen et al., Nature Aging 2022) demonstrated that original chronological-age clocks (e.g., standard Horvath DNAmAge) suffered from substantial technical noise (up to 9 years of deviation between replicates). When noise-reduced PC clocks or high-reliability metrics were applied, the apparent dramatic age reversals reported in small intervention datasets disappeared or were revealed to be false-positive statistical artifacts.

Fact-checked episodes

Publications