Automation and data-driven design of polymer therapeutics

Area of Science:

• Polymer Chemistry
• Materials Science
• Biotechnology

Background:

• Polymers offer tunable properties for drug delivery and biomaterials.
• Traditional polymerization methods (CLRP, ROP) require stringent conditions.
• High-throughput synthesis and screening (HTS) are becoming feasible with automation.

Purpose of the Study:

• To review the impact of machine learning (ML) and artificial intelligence (AI) on polymer therapeutics.
• To highlight data-driven approaches for efficient polymer design.
• To explore applications in drug delivery, gene delivery, and bioactive polymers.

Main Methods:

• Review of state-of-the-art research in polymer therapeutics and data-driven design.
• Analysis of ML/AI applications in small molecule drug discovery and their emergence in drug delivery.
• Examination of advancements in polymerization techniques compatible with HTS.

Main Results:

• ML and AI are crucial for maximizing experimental efficiency in HTS, avoiding "fishing expeditions."
• Data-driven approaches are essential for establishing quantitative structure-property relationships (QSPRs).
• Closed-loop design-build-test-learn workflows are valuable for polymer therapeutics development.

Conclusions:

• The polymer therapeutics field is poised for significant advancements through ML and AI.
• Researchers can now explore the polymer structural landscape more comprehensively.
• Establishing biologically significant QSPRs will accelerate the development of novel polymer-based therapies.