This invention includes an artificial intelligence-based drug design platform to enhance drug design and development. Various computational models are developed for: (1) Hit identification; (2) Lead optimization; (3) Cytotoxicity prediction; and (4) ADME prediction. Our platform has identified 62 novel inhibitors targeting 36 kinases, 10 of which are active at the nM range.

The innovative feature of this drug design platform combines artificial intelligence with (1) protein structures, (2) compound properties, and (3) molecular interactions to establish models for different drug development stages. Models show increased accuracy to hit identification, cytotoxicity and ADME prediction. The high accuracy of our platform reduces the time and cost for drug development.

On average, new drug development costs $2.6 billion USD over 14 years. The low success rate, long development time and high cost are all impediments for small-molecule drug development. Our drug design platform provides models to accelerate different stages of drug development, reduce costs and increases the success rate for new drugs and is of great potential for the pharmaceutical industry.