Back in 2023 we praised the achievements of Google Deep Mind and its Alphafold system for predicting protein structures (LSS 23 2 23) The step change in productivity , (no disrespect to human biochemists) was so remarkable that we compared it to the invention of stone tools. Since when it has pretty much become a standard tool in medical research.
So much for proteins. What about RNA? Advances in understanding its structure, maybe even making a little of our own, might convey enormous medical benefits. Read this from Nature Briefings Seeking an alphafold moment for RNA
Protein-structure-prediction tools such as AlphaFold have transformed biology. But RNA is a tougher nut to crack: it poses unique molecular challenges, and relatively few data are available to train computational models. So researchers have been getting creative, building a toolkit to aid the prediction of RNA structure that incorporates the latest developments in artificial intelligence.Nature | 10 min read
“A tough nut to crack” Indeed. For one thing RNA has always suffered from that “middle child” syndrome, lost between its more glamorous siblings, DNA and proteins. So there is a lot less data to feed into the AIs. And even the main forms, t-RNA and m-RNA are fiendishly complicated, like any biological macromolecule. Fortunately, there is a superb article from the main part of Nature by Diana Kwon[1] which lays out the problems and challenges with great clarity; well worth a glance, However the advantage of getting on top of RNA and bringing it, so to speak, into the twenty first century could be colossal, Never forget that it was an m-RNA vaccine that finally got the SARS-Cov-2 virus on the run. That is a glimpse of what might one day be acheived.
#rna #AI #alphafold #medical research #biotechnology #nucleic acids #proteins #vaccines
Learning, Science and Society 