Artificial Intelligence Driven Workflow for Accelerating Design of Novel Photosensitizers
🤖 AI Summary
Traditional trial-and-error approaches to photosensitizer (PS) discovery suffer from low efficiency and high cost. To address this, we propose AAPSI—an AI-accelerated PS innovation closed loop—that integrates expert knowledge, scaffold-driven molecular generation, and Bayesian optimization to jointly optimize structural novelty and synthetic feasibility. Methodologically, we develop a graph-transformer model to accurately predict singlet oxygen quantum yield (ΦΔ) and absorption wavelength (λₐbₛ), coupled with a curated database of >100,000 PS–solvent pairs for intelligent virtual screening. The workflow generated 6,148 synthetically accessible candidates and identified multiple high-performance PSs; notably, HB4Ph achieves ΦΔ = 0.85 and λₐbₛ = 650 nm—performance competitive with state-of-the-art PSs. This framework significantly enhances the rational design efficiency of PSs and establishes a new paradigm for the development of optoelectronic functional materials.
Application Category
📝 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 extbf{A}I- extbf{A}ccelerated extbf{P}hoto extbf{S}ensitizer extbf{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 ($φ_Δ$) and absorption maxima ($λ_{max}$), 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 ($φ_Δ$=0.85, $λ_{max}$=650nm).
Problem
Research questions and friction points this paper is trying to address.
Accelerating photosensitizer design using AI-driven workflow integration
Overcoming trial-and-error limitations in photosensitizer discovery process
Optimizing photodynamic therapy candidates for quantum yield and absorption
Innovation
Methods, ideas, or system contributions that make the work stand out.
Closed-loop workflow integrating expert knowledge and AI
Scaffold-driven generation ensuring novelty and feasibility
Graph transformers predicting quantum yield and absorption
🔎 Similar Papers
No similar papers found.
