Matt Kaeberlein, PhD

Epigenetic Aging Clocks Lack Reliable Real‑world Accuracy

This is a thoughtful essay on a new preprint from Raghav Sehgal and Albert Higgins-Chen that’s worth your time. It highlights something we don’t talk about enough: for biological aging clocks to be useful outside of research, they need to be both technically reliable and biologically meaningful. https://t.co/IDXoA4ilSn The key finding: many epigenetic clocks are technically reliable, but show substantial biological variation—including large swings in estimated biological age over just 24 hours. That raises an interesting question: Are these clocks detecting real, rapid changes in biology… or just noise? For now, let’s set that aside and focus on what this means in practice. Bottom line: their real-world utility today is somewhere between limited and non-existent. Why? Because we don’t know the error bounds. If you can’t quantify the uncertainty of a measurement, you can’t interpret changes. A shift of a few years—or even much larger—may not reflect anything meaningful at all. This preprint helps define the floor of that variability under controlled lab conditions. But the ceiling is almost certainly much higher. In the real world: samples are often collected by consumers via finger prick or spitting in a tube; samples are shipped through the mail where environmental variables like temperature and humidity can vary dramatically; handling and processing is often outsourced to third-party vendors with unknown technical skills and QC standards. Each step adds noise. My best guess: real-world variability is at least 2–3x higher than what’s reported in the pre-print. That doesn’t mean these tools are useless. It means they need to be used thoughtfully, and we should be honest about what they can—and can’t—tell us today. Right now, they’re much better suited for population-level research than for individual decision-making or clinical application.