QuantaBricks

Decentralize Computer-aided Drug Discovery
Discord:

[Jo @jomyzc]

1/Manuscript accepted today @ACSPublications This March, @tulalanAndLbuta and I used the ChemOrchestra platform https://t.co/jEDyWEItnC to optimize and dock 195 drug-like molecules on pure CPU. #cadd #chemistry #drughunter #drugdesign #aidd

[Jo @jomyzc]

One year after #davidbaker won the #NobelPrize for #alphafold , many companies are still investing heavily in AI protein structure prediction. But in reality, this is just a numbers game — the accuracy of the models isn’t actually that critical. That’s because experimental structures are derived from crystallized or frozen #protein, while proteins in living cells may differ significantly. As long as the prediction error of the folding model is smaller than the experimental error, that’s already sufficient — and most models can achieve that.
What’s more important is that the model is lightweight, has low hardware requirements, and doesn’t rely on MSA servers.
Try ESMfold and pdbfixer on ChemOrchestra,and curious to know how AI protein structure prediction can practically solve your research/industry problems: https://t.co/2ZC8UrDv1W