Ex Scientia: Commercialising Multi-Target Drug Design

一种药物的临床疗效和安全性取决于它在蛋白质组中跨多个蛋白质的活性分布。因此，针对多种蛋白质的多药理图谱进行药物先验合理设计的方法在药物发现中具有巨大的价值。我们已经开发并实验验证了一种新的发现配体的方法，通过进化自动设计化合物。该设计方法已经通过40个新化合物的合成和800个配体-目标预测的分析进行了测试，其中75%被证实是正确的。我们建议将该算法商业化，成为一个创新的药物发现平台。特别是，我们建议将该算法扩展到云计算基础设施上，以确定多靶点药物设计机会的组合。此外，我们建议将算法的验证数据从GPCR扩展到激酶多药理学预测。

The clinical efficacy and safety of a drug is determined by its activity profile across multiple proteins in the proteome. Therefore methods to rationally design drugs a priori against polypharmacology profiles of multiple proteins would have immense value in drug discovery. We have developed and experimentally validated a new approach for the discovery of ligands, by automatically designing compounds by evolution. The design method has been tested with the synthesis of 40 novel compounds and the assaying of 800 ligand-target predictions, of which 75% were confirmed correct. We propose to commercialise the algorithm into an innovative drug discovery platform. In particular we propose to scale the algorithm onto a cloud computing infrastructure to identify a portfolio of multi-target drug design opportunities. In addition we propose to extend the validation data of the algorithm from GPCR to kinase polypharmacology predictions.