Quantum-assisted fragment-based automated structure generator (QFASG) for small molecule design: an in vitro study

<b>Introduction:</b> The significance of automated drug design using virtual generative models has steadily grown in recent years. While deep learning-driven solutions have received growing attention, only a few modern AI-assisted generative chemistry platforms have demonstrated the ability to produce valuable structures. At the same time, virtual fragment-based drug design, which was previously less popular due to the high computational costs, has become more attractive with the development of new chemoinformatic techniques and powerful computing technologies. <b>Methods:</b> We developed Quantum-assisted Fragment-based Automated Structure Generator (QFASG), a fully automated algorithm designed to construct ligands for a target protein using a library of molecular fragments. QFASG was applied to generating new structures of CAMKK2 and ATM inhibitors. <b>Results:</b> New low-micromolar inhibitors of CAMKK2 and ATM were designed using the algorithm. <b>Discussion:</b> These findings highlight the algorithm's potential in designing primary hits for further optimization and showcase the capabilities of QFASG as an effective tool in this field.