A Genetic Anchor for Health express

Original: "Is It You or Your Environment? A Bayesian Inference Framework for Genomically-Anchored Personalized Physiological Interpretation"

· Aruna Dey, Suraj Biswas

arXiv:2606.13556 · 2026-06-11 · CC BY 4.0 · 2 min

Artificial Intelligence cs.HC q-bio.BM q-bio.GN q-bio.MN

Your genome could become the baseline for telling a temporary glitch from a permanent trait.

You buy a new kitchen scale, but instead of exact weight, it only shows deviations from some "norm." To find out how much an apple weighs, you first have to weigh dozens of other apples and calculate the average. That's roughly how today's health monitoring systems work: they spend weeks collecting data on heart rate, blood pressure, and sleep to figure out what's usual for you and what's an anomaly. This is the cold-start problem — the model doesn't know you until it accumulates enough observations.

In a new study, scientists propose not to wait for weeks but to give the system a "cheat sheet" right away — your genetic profile. Genes don't change with stress or lack of sleep; they're like the foundation of a house, laid during construction. The researchers build a personal baseline on this foundation: they take hundreds of known genetic variants (the ones that large studies link to blood pressure, heart rhythm, metabolism) and calculate what level of heart rate or blood sugar is "programmed" specifically for you. This isn't a rigid verdict but rather an initial hypothesis that gets refined with each new measurement.

A heart rate of 55 beats per minute for someone with a genetic predisposition to a high heart rate would signal relaxation, while for someone whose genes predict 80 beats, it would be a sign of depression or illness. Without a personal anchor, these two situations are indistinguishable.

As real data comes in, the genetic "cheat sheet" gradually loses weight, giving way to live observations. The formula is simple: initially, the system relies almost entirely on the genetic prediction, and after a couple of weeks, it relies on what the fitness tracker has measured. But even at the end, the genetic anchor doesn't disappear completely: it's like an old friend who has known you since birth and helps you not to mistake a temporary cold for a chronic condition. This approach has already been tested for six physiological indicators, from heart rhythm to blood sugar levels, and uses only the most reliable, repeatedly verified genes, not dubious "aggression genes" or "happiness genes."

A genetic anchor doesn't predict the future; it sets a coordinate system. It's like knowing your house is built on a hill: you'll measure any changes in water level not from the "average height in the area" but from your own foundation.

Whole graph →

The difference between genetically predicted and actual heart rate can tell you more about stress than the heart rate itself: it's like a scale showing not just "70 kg" but "plus 2 kg after dinner."

In the series "Altered Carbon," a personality can be transferred to a new body, but a genetic anchor would make such technology even more complex: your "normal" is tied to DNA, not consciousness.

\hat{G} = \mu + \sum_i \beta_i g_i

Genetic prediction of a physiological indicator: \mu is the population mean, \beta_i is the effect size of the genetic variant, and g_i is the number of copies of the variant in an individual.

Scientists

Andrey Markov

Concepts

personalized medicine prior distribution ai self correction genetic algorithm bayesian hierarchical model predictive model frame granger causality Bayesian inference genetic code ai speech analysis prediction error benchmark prediction diagnosis Haeckel's biogenetic law Braess's paradox in biology bioinformatics Fisher's fundamental theorem of natural selection Generative Adversarial Networks

I am the author

A Genetic Anchor for Health express