AI-Powered Knowledge Base Enables Transparent Prediction of Nanozyme Multiple Catalytic Activity

The DiZyme - web-resource for rational design of nanozymes using machine learning algorithms.
It contains a unique expandable database of nanozymes with links to original articles, an interactive clickable tool for its visualization, and a machine learning models for various levels of user requests (explorative, detailed and customised) capable of predicting multiple catalytic activity represented as the Michaelis-Menten (Km, mM) constant with R² 0.75 and the maximum reaction rate (Vmax, mM/s) with R² 0.77.

Nanozymes are a unique class of materials that have already demonstrated a number of useful properties for applications in biomedicine, biosensing, clinical diagnostics, environmental monitoring and beyond. At the same time, the search for new candidates is usually associated with a large amount of tedious experimental work. To simplify experimental screening, we were the first to develop a more convenient, machine learning (ML) driven approach to search for new candidates and released the web platform, called DiZyme.

Citation

Julia Razlivina, Andrei Dmitrenko, Vladimir Vinogradov. AI-Powered Knowledge Base Enables Transparent Prediction of Nanozyme Multiple Catalytic Activity. J. Phys. Chem. Lett. 2024, 15, XXX, 5804–5813. DOI: 10.1021/acs.jpclett.4c00959