Learning objectives

1. Describe how graph neural networks predict antibacterial activity and evaluate toxicity to identify selective antibiotic candidates.
2. Explain how generative deep learning explores uncharted regions of chemical space to design structurally novel antibiotics.
3. Understand how AI-designed molecules are validated through experimental testing and mechanistic studies.

Speaker biography

Aarti Krishnan is a postdoctoral researcher at MIT specializing in AI-driven drug discovery for infectious diseases. To combat multidrug-resistant bacteria, her recent work leverages deep learning and generative AI approaches to design structurally novel antibiotics by exploring previously inaccessible chemical space (Krishnan et al. Cell, 2025). Her interdisciplinary research bridges computational and experimental biology and extends to parasitology, where she investigates metabolic pathways in eukaryotic pathogens to identify and target new therapeutic vulnerabilities.