Report comment

I personally don't agree with this view. Structure-based drug design (SBDD) campaigns greatly help in reducing the failure rate (without SBDD, it'd be likely higher than the one mentioned in the article). Of course, AlphaFold's models are still predictions, but in the absence of experimental structures, and coupled with other computational methods (e.g. MD simulations), it's definitely much better than just trying our luck by pouring a chemical library onto cells/purified proteins (as it is still often done, for various reasons). That approach makes more sense after designing e.g. a scaffold within a protein structure/model, doing some medicinal chemistry work on it, synthesizing these compounds, and then performing said enzymatic/cellular assays.