Machine Learning-driven Lead Optimization in Anticancer Drug Discovery
DescriptionMore than 99% of lead compounds discovered in anticancer drug screening, which requires billions of dollars to develop, failed in clinical trials. A major bottleneck in this process is lead optimisation, synthesis of structural variants of drug candidates to maximise efficacy and biosafety. Artificial intelligence is widely used in anticancer drug development, but most machine learning (ML) algorithms are developed to predict dose responses and mechanisms of action of lead compounds. Our team proposes to use ML in lead optimisation. Databases on in vitro cytotoxicity, solubility and in vivo toxicology will be used to train algorithms that canpredict molecular features for improving cancer selectivity, pharmacokinetic and biosafety properties. The use of ML in lead optimisation can potentially speed up the preclinical phase of drug discovery. Using lead compounds discovered by our team as prototypes, we will assess and optimise the effectiveness of this approach for future drug screening projects.
|Effective start/end date||30/06/21 → …|