[Colloquium] Priyanka Setty MS Presentation/Mar 28, 2024

[Megan Woodward]

This is an announcement of Priyanka Setty's MS Presentation
===============================================
Candidate: Priyanka Setty

Date: Thursday, March 28, 2024

Time: 10 am CT

Location: JCL 298

Title: AMP Architect: A Multi-Agent AI Framework for Antimicrobial Peptide Design

Abstract: In response to the escalating crisis of antimicrobial resistance (AMR), which accounts for an estimated 4.95 million deaths globally each year, we propose to develop a self-driving laboratory for designing antimicrobial peptides (AMPs). AMPs are small, naturally occurring molecules recognized for their broad-spectrum antibacterial activity and unique mechanisms of action, positioning them as promising candidates for the development of new therapeutics against drug-resistant bacteria.

Conventional in vitro and in vivo experimental assays, while considered the benchmark for AMP discovery and validation, are limited by the immense combinatorial complexity of potential AMP sequences. For instance, designing, a peptide of merely 20 amino acids in length presents an astronomical 20^20 〖^〗possible combinations, rendering traditional methods insufficient. To autonomously navigate this vast space, we propose the development of AMP Architect, a novel multi-agent AI framework specifically designed to generate, screen, and validate potential AMP candidates. This system integrates and analyzes data from diverse sources, identifies peptides with specific functions, and predicts their efficacy and safety profiles through a combined computational and experimental approach.

A core component of this self-driving laboratory is the integration of generative AI with state-of-the-art robotic platforms, which autonomously execute growth assays to assess the antimicrobial efficacy of AMPs. This approach not only expedites the discovery process but also enhances the precision of screening high-potential AMP candidates. By amalgamating these technological advancements within the multi-agent framework, we envision such a multi-agent workflow will enable a transformative step in our ability to combat AMR and reduce  the global health burden it currently imposes.

Advisors: Ian Foster

Committee Members: Rick Stevens, Arvind Ramanathan, and Ian Foster



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20240327/f5a7b19f/attachment.html>