Thread by @EpgntxEinstein

Andy Feinberg has written a review about #epigenetics in @NEJM.

Unfortunately, it’s damaging to the field of human epigenetics research.

The Key Role of Epigenetics in Human Disease Prevention and Mitigation
nejm.org/doi/full/10.10…

A couple of historical inaccuracies to start. Waddington was saying the opposite – cell fate (the cellular event in epigenesis) is probabilistic, not deterministic. The major point he was making was that genes have a role in influencing epigenesis, a major argument at the time

Waddington did not ignore environment at all. In fact, he treated Drosophila larvae with heat/ether to see whether he could influence cell fate. He considered the epigenetic landscape (cell fate) potentially modifiable by environment and genetic variation, see next image

In attempting to explain the Baldwin effect, Waddington showed this variation of creode depths, suggesting cell fates/epigenesis could be influenced by genetic or environmental variation

In the section “Forms of Epigenetic Information”, the definition of epigenetics=transcriptional regulatory information, equated with DNA methylation and chomatin variation…

…However, this raises the same problem as similar epigenetics reviews over the last 20 years. How do these modifications know where to go? The role of TFs is omitted.

In the section discussing paradigms of environmental effects on the “epigenome”, the studies cited are potentially confounded by cell subtype/functional variant effects. We just can’t draw the definite conclusions being made here.

In addition, many of these changes (due to the environment) could be due to effects on cell fate, getting back to Waddington, the #polycreodism model, there may not be transcriptional regulatory changes in individual canonical cell types.

Depending on your definition of “common epigenetic disease”, cancer is probably an awful paradigm. The mutational landscape could be causing changes in meQTLs, CIMP is secondary to BRAF/IDH mutations, it’s a genetic disease with transcriptional regulatory consequences

The epigenetic epidemiology/EWAS section manages to ignore completely the interpretability problems inherent to these studies. This does the field a massive disservice, enough uninterpretable studies are being funded/published.

Our review is one of many nature.com/articles/nrg.2…

As regards precision epigenetic medicine, the implication appears to be that we can target epigenetic mediators, although this section is kind of vague.

However, if the changes in DNA methylation seen associated with a disease are due to altered TF binding, or cell fates, or meQTLs, it is unlikely that targeting DNA methylation, the readout of these events, will do anything of value.

It’s very disappointing. As a field, nowadays we know better than to make many of the assumptions in this review uncritically.

This review from a major voice in our field in the NEJM will prompt poorly designed, uninterpretable studies. Not a good day for #epigenetics research