A study published in Science Advances conducted virtual screens to identify potential TAAR1 agonists for neuropsychiatric conditions using AlphaFold and homology modeling techniques. AlphaFold models outperformed homology models in virtual screening, leading to the discovery of potent TAAR1 agonists. Compound 65 showed high potency, selectivity, and favorable pharmacokinetic properties. However, AlphaFold struggled with larger synthetic ligands and predicting dynamic protein conformations. Experimental cryo-EM structures provided better insights for complex ligands. The study highlights the potential of machine learning-predicted structures for drug discovery but emphasizes the need for further refinement to accurately predict GPCR-ligand interactions. Compound 65 shows promise as a candidate for developing treatments for neuropsychiatric disorders.
Source link
AlphaFold accelerates discovery of potential antipsychotic drugs by outperforming traditional methods
