Epigenetic aging clock DeepMAge has 3-year error margin and could be the most accurate DNA methylation clock to date
Eleanor Garth
Dec. 9, 2020 (First Longevity) — Deep Longevity, the Hong-Kong based Longevity start-up has released scientific details on its new aging clock. DeepMAge was trained to predict human age on more than 6000 DNA methylation profiles, analysing methylation patterns to estimate age within a 2.77-year error margin, which the company claims is more accurate than any other human aging clock [1].
Longevity.Technology: Aging clocks determine biological age, rather than chronological – how old our bodies are biologically speaking – by looking at biomarkers and assessing therapeutic, genetic, environmental or nutritional impacts on the aging process. Using clocks means researchers can investigate drugs and therapies, rather than waiting to for subjects to die in order to see if lifespan was affected. DeepMAge shows association with a range of age-related conditions, as well as cancer, dementia and obesity, so this research bodes well for preventing and delaying these diseases, rather than trying to cure or ameliorate them.
DeepMAge is a neural network regressor, which was trained on 4,930 blood DNA methylation profiles from 17 studies. It shows biological relevance by assigning a higher predicted age to people with various health-related conditions, such as ovarian cancer, irritable bowel diseases and multiple sclerosis.
[DeepMAge was trained in 4930 blood samples and verified in 1293 samples from 15 independent studies. The median absolute error achieved by DeepMAge is lower than in any other comparable aging clock] Source: Deep Longevity LimitedDNA methylation is an epigenetic regulatory mechanism that switches genetic expression on or off. As time passes and we age, the number of methyl groups added to our DNA increases, with the result that genetic instructions can become garbled, with genes that should be switched on are actually switched off and vice versa. Methylation aging clocks evaluate the numbers of added methyl groups to calculate biological age.
People with certain health conditions are evaluated by the clock as having a biological age that is higher than their chronological age. This means that early diagnostic tools could be developed to use this difference as an early warning sign or to tackle this faster pace of aging in a therapeutic way. Being able to simulate the aging process at the molecular level would mean that drug targeting would be enhanced as the model would reduce the need for live subjects for tests.
[DNAm is an epigenetic mark that regulates gene expression. This mechanism ensures the tissue and functional identity of cells] Source: Deep Longevity LimitedNeural networks are flexible, meaning the architecture can be adjusted to show the effects of calorie restriction, supplements, exercise, or other lifestyle or environmental changes.
Deep Longevity CEO Alex Zhavoronkov said: “We have created DeepMAge to help researchers, health professionals and consumers track aging processes in the organism. As the global population grows older with each generation, the problem of extending the productive period of human lives is getting increasingly important. Solving it is impossible without a way to quantify aging and digitally test our ideas on how to affect it with medicine [2].”
The paper details how DeepMAge predicts that women with ovarian cancer are, on average, biologically 1.7 years older than healthy women of the same chronological age. Similarly, people suffering from multiple sclerosis patients are predicted by the clock to be 2.1 years older biologically. Results were been obtained for several other conditions, including irritable bowel diseases, dementia and obesity.
This raises a chicken-and-egg dilemma: do certain conditions accelerate aging, leaving epigenetic evidence behind that the aging clock detects or does an advanced aging rate increase the risk of certain conditions? Either way, the links between epigenetics and Longevity is something the researchers plan to investigate further through the DeepMAge clock.
[DNAm aging clocks are statistical models that learn to tell young and old people apart based on their DNAm levels at specific CpG sites]. Source: Deep Longevity LimitedFirst author on the paper, which was published in Aging and Disease, Fedor Galkin, Project Manager at Deep Longevity, commented: “Aging clocks have come a long way since the first works by Horvath and Hannum in 2013. We are happy to contribute to this research field. Now, we are going to explore how epigenetic aging can be slowed down with the interventions available to consumers [3].”
Aging clocks are key in both personalised and precision medicine. Deep Longevity is spearheading their development, so we reached out to Polina Mamoshina, DPhil, Chief Scientific Officer of Deep Longevity to get her take on what this new methylation clock means for them.
“At Deep Longevity we are working on a really broad portfolio of aging clocks, ranging from simple blood test based to sophisticated omics like Deep Methylation clocks or DeepMAge,” she told us. “Among molecular markers, epigenetic clocks are far more superior in terms of accuracy of prediction of chronological age. At the same time, its superior accuracy has a downside. Epigenetic profiles are rather stable. That is why it is possible to predict the age of individuals with such good accuracy as the low variation is observed between individuals.
“But we should assume, that epigenetic profiles will be less responsive to interventions unless those interventions directly affect them. That said, we believe that such methylation clocks would be invaluable in the estimation of long-term risks.”
[1] https://doi.org/10.1016/j.arr.2020.101050
[2] https://medium.com/longevity-algos/deepmage-most-accurate-aging-clock-6cce9ca8150b
[3] https://www.eurekalert.org/pub_releases/2020-12/dll-dlp120720.php
.