Wow. Iva’s paper is finally published. There is only one human proteome, and there are many disordered protein regions in there. For the first time systematic predictions of their biological functions are now available. Of course, this is only the *first* attempt to systematically map the function of the human …
Ami’s paper is now published! A protein language model has learned that proteins are made of functional units, often known as protein domains or regions. 10,000s of protein domains have been identified over the years by some of our favourite bioinformatics teams (e.g., pfam). But Ami shows that by looking …
Genome sequences specify 1000s of the components of cells and organisms, usually using millions or billions of A,C,G and Ts. As computational biologists, ideally, we would use the entire genome sequence to make predictions about phenotype. Leveraging recent IT breakthroughs (thanks NVIDIA), Cameron figured out a way train deep learning …
The origin of life is one of the great mysteries of science. Of course, no one was there to observe it and there are many reasons that it’s tricky to figure out how you could get a living thing from non-living material. Inspired by recent research in biological condensates, Alan …
Check out our review paper on large language models for the genome.
Check out our latest idea for how to visualize subcellular localization changes for 1000s of proteins! The map and arrows are made with no training data about localization — directly from the images! (colours are from visual inspection of images 20 years ago.)
Andrew’s paper shows that homologs can be used to augment sequence data
