Designing Enzymes with AlphaProtein Novo (www.biorxiv.org)

A significant breakthrough in enzyme design has been announced with the introduction of AlphaProtein Novo (AP Novo), a machine-learning pipeline that successfully creates enzymes de novo, surpassing traditional methods that rely on existing natural proteins. This innovative approach allows for the tailored design of enzymes for complex reactions, notably achieving high selectivity in synthesizing the pharmacophore piperidine and effectively degrading the environmental toxin DEHP under conditions unfavorable for natural enzymes. The implications for the AI/ML community are profound, as AP Novo demonstrates the capability of machine learning, particularly through the use of AlphaFold 3 predictions and sequence-ensemble-based metrics, to enhance enzyme design. This not only broadens the potential for discovering novel catalysts across various applications but also positions de novo design as a vital complement to natural enzyme diversity. The success in achieving state-of-the-art catalytic efficiencies on established reactions indicates a transformative shift in molecular sciences, paving the way for innovative solutions in both pharmaceutical development and environmental remediation.