AI Derivation and Exploration of Antibiotic Class Spaces

ICLR 2025 Conference Submission12270 Authors

Keywords: fragment-based drug design, antibiotic resistance, pharmacokinetics, hybrid antibiotics, in silico analysis, retrosynthesis, chemical space exploration, machine learning, antibiotic discovery, protein targets

TL;DR: We present an AI-driven tool for predicting pharmacokinetic properties, demonstrated through experiments in historical antibiotic synthesis, hybridization of antibiotic classes, and exploration of new chemical spaces to combat antibiotic resistance.

Abstract: This paper presents a novel approach to fragment-based antibiotic drug design design. We introduce a tool called FILTER, which uses chemical structure data, pathway information, and protein targets to predict pharmacokinetic properties of existing and novel drugs. We report on three distinct experiments utilizing FILTER. The first experiment is an in silico analysis that recreates the historical discovery of penicillin derivatives, validating our approach against known outcomes. The second experiment explores the combination of functional groups from different antibiotic classes to create molecules with multiple mechanisms of action. We refer to this approach as hybridization as all synthesized molecules are composed of fragments from both classes. Our final experiment is forward-looking as it explores new chemical spaces to build a library of promising compounds for further antibiotic development. Throughout all these experiments, FILTER serves as an indispensable oracle, predicting physical properties and potential therapeutic efficacy of the new molecular architectures, aiming to accelerate the drug development process and address the challenge of antibiotic resistance. Our approach represents an ongoing, significant shift from traditional drug discovery methods, emphasizing the role of innovative technologies in combating the urgent global threat of antimicrobial resistance.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 12270