I'm happy to see my latest preprint out on stratigraphic intervals, a topic I've been interested in since my first year as PhD student. Available in the brand-new StratIntervals.jl (github.com/gaballench/S...) @julialang.org package. Preprint on @biorxiv-evobio.bsky.social doi.org/10.1101/2025...

Honoured to have won the prize to best presentation in methods and software at the Congresso Brasileiro de Biologia Evolutiva

How do the reconstructions in the Solesmes edition compare with our Bayesian reconstructions? They tend to be more similar to Frankish nodes, and not necessarily similar to reconstructions at the root, thus apparently representing local traditions rather than general ones.

Reconstruction of chant melodies is, however, available from a more traditional musicological approach: The Solesmes edition of chant melodies. They compiled a massive number of sources and applied musicological criteria for reconstructing the original melodies for use in offices.

Then we used the melody alignments and ancestral melody reconstruction methods for inferring the posterior probability of each possible state at the nodes, thus being able to pick the melody with the highest posterior probability. The one at the root would then be the closest to what the dove sung.

Now we have a rooted tree, and the divergence time estimation provides information on the time of splits in the tree. The topology remains quite similar to our 2023 study, but now giving a clear interpretation of cultural transmission to our hypothesis.

Here we applied Bayesian model selection by rooting at each of the branches in the tree and then calculating the marginal likelihood, which allows us to calculate the model posterior probability and choose the best position for the root.

The results of this first phylogenetic analysis using chant melodies were published in 2023 (archives.ismir.net/ismir2023/pa...). The phylogenetic tree offers insights into the effect of religious context and geographic provenance of these manuscripts.

But how to infer the evolutionary relationships among sources? We use the fact that melodies are encoded like a sequence of pitches, pretty much the same as DNA is encoded as a sequence of nucleotides. We use an evolutionary model to estimate an evolutionary tree of these sources.

Through history, we have numerous manuscripts documenting these melodies in musical notation systems evolving in order to improve the accuracy of melodic encoding. An enormous effort has been invested in compiling these melodies in digital format.

A myth around Gregory's standardisation effort says that he received divine inspiration in the form of a dove, which sung and dictated the melodies to him. He then wrote them to be performed exactly like that across Europe.

But what is Gregorian chant? Music has been involved in religious contexts long ago, and Pope Gregory attempted standardise the melodies which were sung during the religious office throughout Europe, a vast area during Medieval times.

Back in 2022, the Genome of Melody project started aiming to understand the evolution of Gregorian chant using tools from bioinformatics and a considerable amount of melody data encoded from original sources such as manuscripts from all over Europe.

Culture, a complex set of phenomena associated with our species, has been studied extensively using different approaches, especially in the humanities.