Combining yeast chemical genetics and AI to enable efficient identification of molecules from plants and fungi with cell inhibitory modes-of-action re

大量活的和保存下来的植物和真菌收藏，例如在Kew发现的植物和真菌，是药物发现的极其宝贵的资源，但从自然界中识别具有药理选择性的分子显示出理想的作用模式目前既耗时又低效。天然产物化学需要新的有效方法来释放植物和真菌化合物的药用潜力，并确保只有具有示范药理活性的化合物才能进行化学分离。该项目的目的是测试人工智能化学信息学平台的能力，以正确预测所有已发表的植物和真菌代谢物中的哪些化合物，选择性地与抗癌药物靶标mTOR和HSP90以及帕金森病药物靶标--突触核蛋白相互作用。通过将快速电子筛查与这些药物靶标的简单、高度诊断的酵母菌检测相结合，我们的目标是展示一种方法，以绕过作为天然产品领域当前标志的低效实验方法。

Large living and preserved collections of plants and fungi, such as those found at Kew, represent an extremely valuable resource for drug discovery, but identifying pharmacologically selective molecules from nature that exhibit desirable modes-of-action is currently both time consuming and inefficient. Natural products chemistry needs new efficient approaches to unlock the medicinal potential of plant and fungal compounds and ensure that only compounds with exemplary pharmacological activity undergo chemical isolation. The aim of this project is to test the ability of an artificial intelligence cheminformatics platform to correctly predict which compounds out of all published plant and fungal metabolites, interact selectively with the cancer drug targets mTOR and HSP90 and the Parkinson's disease drug target -synuclein. By combining rapid in silico screening with simple, highly diagnostic yeast assays for these drug targets, we aim to demonstrate a way to sidestep the inefficient experimental approaches that are a current hallmark of the natural products field.