Your Sleep Tracker's REM Minutes Are Partly Inherited: Meet MEIS1

You slept eight hours. Your friend slept eight hours. In the morning you compare charts, and yours says you spent a good deal less of the night in REM than theirs did. Same bedtime, same alarm, different night.

Some of that is the coffee at four in the afternoon and the phone at midnight. But some of it, a study showed this year, was settled before either of you was born.

80,013 people wearing a device

Most of what is known about the genetics of sleep comes from asking people how they sleep, and people are not good witnesses to their own sleep. Portas et al. 2026 measured it instead. 80,013 UK Biobank participants wore a device, and from it the researchers took four things: how long each person slept, how efficiently, and how much of the night was REM sleep and how much was non-REM.

A caution that applies to the study and to the tracker on your wrist alike: a worn device works out sleep stages from movement. A sleep laboratory reads them from brain waves. The device's answer is an estimate — a good enough one to find genes with, as it turned out.

One gene, pulling both ways

The study found 20 places in the genome associated with these measures, 12 of them not reported before. The one that stands out is in a gene called MEIS1. It showed strong effects on REM and on non-REM sleep — in opposite directions. In the GWAS Catalog's record of the study, one version of a position in MEIS1 went with about 9 more minutes of REM sleep a night and about 7 fewer minutes of non-REM sleep.

So it is not a gene for sleeping more or sleeping less. It looks more like a hand on the dial between the two kinds of sleep, which is how the authors read it: a role in the balance of REM and non-REM. MEIS1 is also a gene that tolerates damage badly — versions that stop it working are rare in the population, a sign that the body needs it.

Different in women and in men

When the researchers analysed women and men separately, different places turned up: FOXP2 and NRXN3 in women, and LRP1B, NPBWR2 and PABPC4 in men. The signals as a whole pointed at two parts of the brain, the hypothalamus and the frontal cortex. And a method that uses genes to test cause and effect supported a link between shorter sleep and higher risk of heart and metabolic disease.

Where this touches this site

The study's main MEIS1 position is not in this catalogue yet. Four of its other positions are: rs2006810 in AUTS2, which the study reports for sleep efficiency; rs62033400 in FTO, reported for REM sleep and filed on this site under body mass index; rs62158211, reported for REM sleep; and rs6940638, reported for sleep duration.

MEIS1 itself is no stranger here. A different position in the gene, rs2300478, is on this site's restless legs syndrome page — the urge to move the legs that gets worse at night and breaks up sleep. The sleep and circadian rhythm and insomnia pages and the sleep topic hold the rest, and the Gene Knowledge Graph shows how one gene reaches several conditions.

So the next time your chart and your friend's disagree, the fair conclusion is not that one of you is sleeping wrong. Nights differ, and part of the difference is inherited.