AI-driven Lead Discovery

Authors: Lawless M, Miller D, Waldman M
Conference: ACS
Keywords: admet predictor, AI-driven compound optimization, AIDD Module, BACE1 inhibitors, de novo design
Software: ADMET Predictor®
Division: PBPK

AI-driven Lead Discovery

  • Computer based de novo drug design background
  • The AIDD Module in ADMET Predictor®
  • Using chemically intelligent SMIRKS transformations to generate molecules
  • Synthetic feasibility assessment
  • Incorporating pharmacokinetic and ADMET liability predictions
  • Penalizing out of scope predictions
  • Using capping values when objective function value is “good enough”
  • Using external programs to compute objective functions
  • Selecting Pareto optimal compounds
  • AIDD module demonstration

By Michael S. Lawless, David Miller, and Marvin Waldman

Presented at ACS Spring 2022, March 20-24 In person & Virtual, San Diego, Ca

Download Slides