Deep Reinforcement Learning in the Search for New Antibacterials

Discovery and development of new antibiotics is becoming increasingly challenging, and in fact declining in the private sector because antibiotic-resistant bacterial infections are constantly evolving. One way to increase the antibiotic discovery rate is by taking advantage of the strides in Artificial Intelligence (AI)-based deep learning models to find new antibiotics that kill bacteria, as well as being stable and non-toxic to humans. We postulate that growth inhibition of Escherichia coli, and a wide spectrum of pathogens can be achieved with the identification of AI-derived compounds. To examine this, we will use cross-pillar methods to: (1) Discover new antibacterial compounds and test for growth inhibition in E. coli and drug-resistant pathogens, as well as (2) Identify essential bacterial gene targets and combinations of synergistic growth-inhibiting AI-derived molecules. Together, the AI driven approach will uncover new compounds with broad-spectrum activity against pathogens, and reveal their essential target pathways for drug development and/or multi-target therapies.

Faculty Supervisor:

Mohan Babu

Student:

Partner:

99andBeyond Inc.

Discipline:

Life Sciences

Sector:

Professional, scientific and technical services

University:

University of Regina

Program: