“DNA methylation age of human tissues and cell types” (FAQ), Horvath 2013; mainstream overview of results: “Biomarkers and Ageing: The clock-watcher; Biomathematician Steve Horvath has discovered a strikingly accurate way to measure human Ageing through epigenetic signatures”
I printed out the original paper and read it with some labmates.
Very interesting. The age marker he has created is a simple linear combination of the methylation ratio of several hundred CpG sites (places that a class of methylating enzymes act upon in animals) from large public datasets. Some are positively correlated with age and some are negatively correlated.
I would be interested in people trying to decompose it into subsets of CpGs that have most of their change over childhood or adolescence versus those that change constantly or change only after adolescence.
It’s interesting that muscle tissue and adipose tissue shows very poor correlation while blood and epithelium (two cell types which are constantly proliferating) and brain tissue (very little proliferation at least among the neurons themselves) all show very good correlations. The finding that tumors with few mutations showed major age acceleration while those with many mutations showed less is interesting and provides several possible models of what this could mean.
He proposes a model that methylation age represents the cumulative buildup of the results of an epigenetic maintenance system, but at this early date I would not trust any mechanism Ideas just yet. It leaves open the question if this is a biomarker for a functionally significant epigenetic state, or just a marker for time since cell diffferentiation uncorrelated to other functional differences—though cancer was generally associated with older DNA methylation inferred age in the tissue it arose from suggesting it is at least correlated with something important.
I printed out the original paper and read it with some labmates.
Very interesting. The age marker he has created is a simple linear combination of the methylation ratio of several hundred CpG sites (places that a class of methylating enzymes act upon in animals) from large public datasets. Some are positively correlated with age and some are negatively correlated.
I would be interested in people trying to decompose it into subsets of CpGs that have most of their change over childhood or adolescence versus those that change constantly or change only after adolescence.
It’s interesting that muscle tissue and adipose tissue shows very poor correlation while blood and epithelium (two cell types which are constantly proliferating) and brain tissue (very little proliferation at least among the neurons themselves) all show very good correlations. The finding that tumors with few mutations showed major age acceleration while those with many mutations showed less is interesting and provides several possible models of what this could mean.
He proposes a model that methylation age represents the cumulative buildup of the results of an epigenetic maintenance system, but at this early date I would not trust any mechanism Ideas just yet. It leaves open the question if this is a biomarker for a functionally significant epigenetic state, or just a marker for time since cell diffferentiation uncorrelated to other functional differences—though cancer was generally associated with older DNA methylation inferred age in the tissue it arose from suggesting it is at least correlated with something important.