We analysed global sequence databases to conclusively show that an antiviral drug called molnupiravir has resulted in viable SARS-CoV-2 viruses with significant numbers of mutations, in some cases with onwards transmission of mutated viruses.

This work, completed as part of my residency at Google AI, uses deep residual networks to predict protein function from amino acid sequences. We show that these networks are able to perform this task effectively, in a way that complements BLAST-based approaches, and that they learn to place protein sequences into a generalised embedding space that facilitates downstream applications. Using TensorFlow JS, we built a tool that performs protein functional inference in the browser, client-side. The paper is presented in an interactive form that allows the reader to explore our work and try the models.

Taxonium is the first tool to allow trees of millions of nodes to be readily explored in the browser.

Here we constructed a model of SARS-CoV-2 genomic epidemiology in the UK during 2020-21, chronicling the rise of first the Alpha lineage and the Delta lineage, using data generated at the Sanger Insitute from the sequencing of positive Pillar 2 tests.

Chronumental is a new tool that allows each node on a phylogenetic tree to be placed in time. Previous tools typically worked only for trees of up to tens of thousands of nodes, but Chronumental readily scales to pandemic-scale trees with millions of nodes.