Technologies in their infancy often suffer from “teething problems,” and in this respect, epigenetic age testing has not been spared – especially since the idiosyncrasies of biology don’t always lend themselves to highly consistent statistical modelling.
Some issues were relatively minor and have already been addressed. But if you’ve been following the scene for the past few years and regularly experimenting with different clocks, you will no doubt have encountered at least one problem – a rather BIG problem, we find!
Today, we won’t go into the fact that different clocks yield different biological age results, since there are good reasons why they do so. Instead, we’d like to dwell on the fact that, until very recently, replication consistency using the same clock represented a major challenge for most biological age tests on the market. You tested yourself on Day 1, getting result X. You then tested yourself again on Day 2 (or even on the evening of Day 1), and you got a huge discrepancy with result Y. How come?
An article published in Nature Ageing in July 2022 (Higgins-Chen et al.) revealed that with six of the most famous clocks, replication testing could yield discrepancies of up to 9 years! Ouch… In a nutshell, this state of affairs was attributed to “technical noise”. Meanwhile,there have been many efforts to e.g. retrain models in order to lessen the noise.
Nevertheless, we find it isn’t good enough – especially if you are intent on monitoring subtle lifestyle changes or new supplementation.
So, here comes our newest epiAge experiment: an anecdotal mini case-study of n=1 (m). Our test person tests his biological age regularly (between once a week and once a fortnight) with epiAge and this is the pattern he obtained over the course of approximately 3 (relatively) uneventful months:
As you can see, the biological age of our test person oscillates between a minimum of 46,78 (on 29/12/2023) and a maximum of 53,72 (on 05/01/2024). If you have a closer look, you will notice that most of the positive or negative outliers (marked in blue) are also associated with lower accuracy scores, such as on the 01/12/2023 where accuracy is down to 96,96 as opposed to the prior score of 98,42. Unfortunately, occasional lower accuracy scores or failed tests (as on the 22/12/2023) are almost unavoidable as sample quality (hence useful DNA) also varies due to various biological and lifestyle variables (such as, e.g. stress, poor sleep, suboptimal diet, hormonal cycles, etc.) or occasionally heightened bacterial/viral load. Actually, the highest outlier (53,72 on 05/01/2024) corresponds with our test person experiencing the flu.
On the whole, testing with epiAge yields remarkably consistent results (and this despite a mean absolute deviation (MAD) potential of 2,8 years), as you can witness it in graph form:
So, you may be left wondering… How does epiAge manage that? Well, it really helps that we focus on 13 CpGs that correlate particularly well with ageing. Moreover, our Next Generation Sequencing (NGS) approach ensures more representative and abundant sequencing of these crucial areas. So, you could say that our clock design motto is “lean but intense”, leading to pretty impressive consistency! This is why so many clinicians and biohackers rely on the epiAge test to monitor their short and long-term interventions.
Why don’t *you* give it a go and monitor your biological age with epiAge to test our premise? Today’s experiment may still be anecdotal, but you may help us prove otherwise. We look forward to hearing from you!
Source
Higgins-Chen, A.T., Thrush, K.L., Wang, Y. et al. A computational solution for bolstering reliability of epigenetic clocks: implications for clinical trials and longitudinal tracking. Nat Aging2, 644–661 (2022). https://doi.org/10.1038/s43587-022-00248-2. Online: https://www.nature.com/articles/s43587-022-00248-2#citeas
Illustration
Felix Mittermeier / pexels & epiAge
Back to all posts
