AlphaFold found thousands of possible psychedelics. Will its predictions help drug discovery?
Nature
|
Contributed by: Reid Stephan
Summary
AlphaFold, DeepMind's AI tool for predicting protein structures, identified hundreds of thousands of new psychedelic molecules, potentially aiding antidepressant development. Despite skepticism and rivalry with experimental methods, AlphaFold's structures proved effective in drug discovery. In one case, they identified potent activators of a serotonin receptor, a target for non-hallucinogenic antidepressants. Predicted and experimental structures yielded completely different drug candidates with similar success rates. Recent advances indicate that AlphaFold may speed up projects and help discover new drugs, although they can't fully replace experiments.
