Machine Learning-Assisted Prediction Of Photothermal Metal-Phenolic Networks

Photothermal therapy (PTT) demonstrates significant potential in cancer treatment, wound healing, and antibacterial therapy, with its efficacy largely depending on the performance of photothermal agents (PTAs). Metal-phenolic network (MPN) materials are ideal PTA candidates due to their low cost, good biocompatibility and excellent ligand-to-metal charge transfer properties. However, not all MPNs exhibit significant photothermal properties, and the vast chemical space of MPNs (over 700,000 potential combinations) complicates the screening of high-photothermal materials. This study aims to apply machine learning (ML) to efficiently screening MPNs with optimal photothermal performance. A database of photothermal properties for 80 MPNs was constructed, and a novel feature descriptor was developed to represent MPN characteristics. Multiple ML models were trained and the XGBoost model was used to screen high photothermal MPNs from a virtual database of 44,438 MPNs. Further screening was conducted based on predicted probabilities and factors such as availability, cost, and delivery time for experimental validation. Representative photothermal MPNs were selected for photothermal antibacterial experiments. The XGBoost model successfully identified 1,654 high-photothermal MPNs, from which thirty MPNs were selected for further experimental validation. Experimental validation confirmed a prediction accuracy of 70%. Several previously unreported high-performance MPNs were discovered, demonstrating strong photothermal antibacterial effects. This study demonstrates an innovative ML-driven approach for efficiently screening MPN materials, accelerating the discovery of high-performance PTAs for PTT and biomedical applications.