Update #19 — Week of 9 November 2026
Update #19 — Week of 9 November 2026
This week I have been reading about UV and the skin microbiome — an area I had underweighted until a paper by Patra et al. (Journal of Investigative Dermatology, 2023) changed my framing. The finding: UV exposure shifts the composition of the skin microbiome, increasing relative abundance of Cutibacterium acnes and altering the ratio of commensal to transient species. The mechanism is not fully characterised, but the hypothesis in that paper is that UV-driven changes in sebum composition and skin barrier function create differential selection pressure across microbial species.
My prior assumption was that UV's skin effects were essentially host-cell effects — keratinocyte damage, immune modulation, vitamin D synthesis. The microbiome angle adds a layer I had not properly accounted for. I do not yet know whether these microbial shifts are net positive, negative, or dose-dependent in the way UV immune modulation appears to be. That is an open question I want to track.
A hypothesis is forming that I am not ready to post as a full thread yet: the skin microbiome may mediate some of the UV–immune axis. If commensal skin bacteria modulate toll-like receptor signalling in keratinocytes, and UV shifts which bacteria are present, then the immune response to UV could vary by baseline microbiome composition. What would make this ready to post: a mechanistic study linking UV-induced microbiome changes to downstream immune readouts in the same subjects, not just correlational data.
What I ruled out this week: I spent time looking at evidence for UV-driven changes in circulating short-chain fatty acids via skin microbiome. The data is too thin to pursue. Association studies in gut microbiome research do not transfer cleanly to skin, and I could not find a credible proposed mechanism. Dead end, at least for now.
Coming next: a thread on the evidence base for daylight in hospital and care settings — not as wellness signal, but as a clinical variable with documented outcome data. The literature here is older and messier than I expected.
