Профиль

We have produced a working hypothesis for future work on random polypeptide libraries of different composition, on ancestrally reconstructed synthetases, and on non-stationary substitution models. The origin of the genetic code is finally an answerable question. 13/13

The retrofunctionalization hypothesis is that late-added amino acids tend to recruit older, less specific synthetases. We find that the signal comes from early amino acids with late-resolved promiscuities using more complex synthetases. 12/13

Again as a control, this didn’t work when using Trifonov’s order. 11/13

We could add amino acids in order at nodes of synthetase chronograms if and only if we hypothesized exactly the same promiscuities that had been independently suggested by the 2-1-3 rule. The trees also make sense re promiscuities, deep mutational scanning data, and codon adjacencies. 10/13

As a control that we weren’t hallucinating patterns, we did the same thing with Trifonov’s (2000) influential hypothesized amino acid ordering. We got no satisfyingly plausible ambiguities, and worse adherence to the 2-1-3 rule. 9/13

Serine’s codons would once have been mutationally connected, before cysteine and arginine took some of them. 8/13

Adding amino acids in order reveals 2 glaring exceptions to the 2-1-3 rule. They can be resolved by assuming ambiguous coding of hydrophobic amino acids V/I/M, and of small/polar amino acids A/T, mirroring exchangeability in deep mutational scanning data, and latent synthetase promiscuity. 7/13

We infer ancestral codes by position amino acids, starting with VTG (left), based on their current codon usage (right), following the 2-1-3 rule. Glycine is so small that its polar backbone is more important – the first code would have a good hydrophobic-polar mix. 6/13

From where similar amino acids are, the 2-1-3 rule posits that information within 3nt codons once came only from the middle position, then the 1st, with the last still not fully used today. Distinguishing purines from pyrimidines came before distinguishing within each. 5/13

Our order doesn't support the coevolution theory that amino acids upstream in metabolic pathways were added first. 6 pairs work (S before C and M, putting E before Q, P, and R, and putting T before I), but 4 don't (D after T, M, and K, with N a tie, and G after S). 4/13

The relationship is stronger still when considered separately for amino acids handled by class I vs class II amino acyl tRNA synthetases, suggesting that it was the evolution of synthetases that drove progression toward larger amino acids. 3/13

Minor improvements to our previous estimate bsky.app/profile/joan... of which amino acids are enriched vs depleted in LUCA (indicating their order of appearance) strengthened the signal that small amino acids came first. 2/13

Linking here also to our thread, originally on the preprint bsky.app/profile/joan...

I'm using that pdf to get AI to build me a spreadsheet. But given weirdnesses of not all talks starting at the same time, I'm only building it for sessions that I might go it rather than all sessions. I'm happy to share my prompt if it helps, and if someone else wants to build a complete one.

They replied that not starring individual talks "was unfortunately the only feasible solution for now". Despite other conferences managing. But a very rough pdf can now be downloaded from "Download session and presentation list" at smbe2026.org/programme/.

Perhaps this is a question of different ways of phrasing things?They are saying that the "mutation effective population size" is different from the coalescent or drift effect population size, being instead more like the census size. This seems aligned?

I agree. Sorry I missed your paper. I does remind me of this one doi.org/10.1371/jour... by @petrovadmitri.bsky.social

Evidence that Adaptation in Drosophila Is Not Limited by Mutation at Single SitesAuthor Summary Adaptation in eukaryotes is often assumed to be limited by the waiting time for adaptive mutations. This is because effective population sizes are relatively small, typically on the ord...doi.org

The result does get weaker when the proportion of adaptive mutations becomes high, but doesn't go away.

Results are surprisingly insensitive to mean selection coefficients. The first attached figure is for the zero environmental change case, the second (top row is what matters) is broader.

Meltdown only becomes more important when beneficial mutations are not so rare compared to deleterious mutations. So population genetics should never ignore beneficials over long timescales: either they are common, or their limiting nature is critical. 5/5