GrimAge: methlyation DNA change, mortality

[jim mcnamara]

TL;DR Summary

DNA methylation - Grimage strongly predicts lifespan and healthspan

Popular precis:
https://www.aging-us.com/dna-methylation-grimage-strongly-predicts-lifespan-and-healthspan Ake Lu, Steve Horvath
David Geffen School of Medicine at UCLA

DNA biomarkers measure an individual’s risk of mortality by analyzing positions on the DNA where methyl groups change with age. These positions are analysed by applying DNA from blood onto a “chip”, which measures the degree of their methylation.

My take:
While DNA may change anywhere due to methylation, so that cells in a given tissue may have different damage points on DNA, there apparently is a Universal "calendar" (or clock if you like) that is consistent for recording methylation changes. Sampling methylation states using ~1000 places on sample DNA from the parts of this calendar is the first step. Next step is applying metrics using GrimAge software on the sample data to give an overall estimate of age-related DNA damage. The changes to DNA start in utero. And persist throughout life.

One result is in units called packyears - number of years of smoking.

Aging appears to be an open source journal. @Ygggdrasil likely knows more details about this universal calendar. I'm strictly an outsider.

 

[mfb]

18% more accurate than calendar age. Does that mean if mortality tables predict 20+-10 years they can predict with an uncertainty of 10/1.18=8.5 years? The original study doesn't have "18" anywhere with such a meaning.

 

[jim mcnamara]

The popular article is based on an interview. It could an explanatory example from that interview...? Merely a guess.

 

[Laroxe]

Methylation is one of the many methods used to control gene expression, in this case it stops gene expression and it is a long term action. Its not a measure of DNA damage which clearly does accumulate with age. I suspect there are a great many processes in our body that are turned on at particular times in our lives and then turned off, its a characteristic of our development. So it would make sense that as we age we would see more evidence of gene expression being disabled, we also accumulate in our tissues an increasing number of senescent cells, cells that are alive but essentially none functioning. My own view is that it is these things that this test might be measuring, really we have little information about specific blood proteins that are reliable indicators of future health, if we did this is what we would be measuring.
I can't see the predictive accuracy will have people clamouring for this test to be available and I think it requires some confirmation.