This discovery science is only possible thanks to the large-scale somatic phylogeny datasets re-analysed here, generated by Emily Mitchell, Stan Ng, Matthias Wilk,Kenichi Yoshida,Jyoti Nangalia+others, funded by @cancerresearchuk.org @sangerinstitute.bsky.social @wellcometrust.bsky.social + others

Intriguingly, blood stem cells had a particular type of very long-lasting damage (~2-3 years), leading to 15-20% of their mutations – some contributing to cancer. This damage wasn’t evident in other tissues. We have theories, but we don’t yet know why.

It was these patterns of unusual mutation inheritance, or multiple different mutations at the same site in closely related cells that was the key to recognizing & characterizing these unusual types of long-lasting damage.

If the base is partially recognizable, the DNA copying machinery may flip between copying it right, and copying it wrong in one specific way. This will only cause 1 mutation, but the pattern of inheritance will not fit a single acquisition event (a ‘phylogeny-violating variant’)

But what if it’s not? If the DNA damage sticks around through multiple rounds of cell division & DNA replication it may be misread in different ways in each round. This will lead to different mistakes at the same position (a ‘multi-allelic variant’).

If the damaged base is present during DNA replication it may be misread, resulting in permanent mutations that can contribute to cancer development. However, the DNA damage itself is usually recognized and mended quickly by repair mechanisms in our cells.

How did we work this out?

DNA damage is distinct from a mutation. While a mutation is one of the 4 standard DNA bases (A, G, T or C) in the wrong place (like a spelling mistake), DNA damage is chemically altered DNA (more like some illegible writing).

I work with somatic phylogenies: family trees of 100s of cells from 1 individual, illustrating their relationships going back to conception. It is inferred from the pattern of shared mutations in the DNA. In 2019 I noticed a mutation that didn’t fit the phylogeny – the 'blip'

So pleased to have the paper out, available here: www.nature.com/articles/s41586-024-08423-8
For me this was discovery science as I had always hoped it would be. A lot of fun, and some proper detective work with plenty of twists & turns on the way. Brief thread below

In science, we often see weird blips in the data. The question: is it artefact (usually!), or something new & exciting? We don’t always have time to dig deep.

Our paper in @nature.com today came from just such a blip. So don’t ignore the weird stuff. Pull on that thread...