Artificial Intelligence Driven Workflow for Accelerating Design of Novel Photosensitizers

TL;DR

This paper introduces AAPSI, an AI-driven closed-loop workflow that combines expert knowledge, scaffold-based molecule generation, and Bayesian optimization to rapidly design novel photosensitizers with high efficiency and structural novelty.

Contribution

The work presents a novel AI-accelerated workflow integrating scaffold-based molecule generation and Bayesian optimization for photosensitizer discovery, demonstrating its effectiveness with experimental validation.

Findings

Generated 6,148 candidate molecules with high synthetic feasibility.

Identified a candidate with $\

Achieved a photosensitizer with $\

Abstract

The discovery of high-performance photosensitizers has long been hindered by the time-consuming and resource-intensive nature of traditional trial-and-error approaches. Here, we present \textbf{A}I-\textbf{A}ccelerated \textbf{P}hoto\textbf{S}ensitizer \textbf{I}nnovation (AAPSI), a closed-loop workflow that integrates expert knowledge, scaffold-based molecule generation, and Bayesian optimization to accelerate the design of novel photosensitizers. The scaffold-driven generation in AAPSI ensures structural novelty and synthetic feasibility, while the iterative AI-experiment loop accelerates the discovery of novel photosensitizers. AAPSI leverages a curated database of 102,534 photosensitizer-solvent pairs and generate 6,148 synthetically accessible candidates. These candidates are screened via graph transformers trained to predict singlet oxygen quantum yield ($\phi_\Delta$) and absorption maxima ($\lambda_{max}$), following experimental validation. This work generates several novel candidates for photodynamic therapy (PDT), among which the hypocrellin-based candidate HB4Ph exhibits exceptional performance at the Pareto frontier of high quantum yield of singlet oxygen and long absorption maxima among current photosensitizers ($\phi_\Delta$=0.85, $\lambda_{max}$=650nm).