In this episode of Mendelspod, we plunge into the shape-shifting world of RNA with Dr. Gavin Knott, associate professor at Monash University and alum of Jennifer Doudna’s lab. Knott is part of a new generation of researchers setting structural biology ablaze with AI, decoding RNA-guided systems and expanding the CRISPR toolbox far beyond its original scope.
Knott introduces the listener to RNA not just as code, but as living language. He discusses how AI now allows scientists to see molecules in 3D at the beginning of a project rather than years into it: “You can take that protein sequence, run it through something like AlphaFold, and there it is, looking at you.”
0:00 RNA - the shapeshifting paradox of a molecule
4:05 Using AI: here’s an idea—maybe we know enough about biology
6:50 The ARC Institute, and that one paper
9:05 Structural biology at everyone’s fingertips
18:10 Why are gene therapy companies going so slow?
23:40 The grammar of RNA
With infectious curiosity, Knott explains how protein design is entering an era of intent, not just discovery. “Maybe we understand enough about the rules of biology,” he posits, “that we can design the proteins we need.” He describes his lab’s twin tracks: mining microbial genomes for novel RNA-guided machines and using generative AI tools like RosettaFold and BindSpace to build bespoke editors of DNA and RNA.
The conversation touches on emerging efforts to mitigate off-target effects in gene editing using AI, the data limitations holding back RNA-protein modeling, and a powerful metaphor comparing static protein images to photos of horses.
Knott makes a striking point: “We’ve known the spelling of RNA for a while, but now we’re trying to learn its grammar.” This, he explains, means understanding not just RNA’s sequence but its shape, movement, and silent punctuation marks—an effort he sees as the next frontier.
Share this post