A team led by Professor Dominik Niopek from the Institute of Pharmacy and Molecular Biotechnology (IPMB) at Heidelberg University has developed a new artificial intelligence model called ProDomino, which can predict the assembly methods of protein subunits, thereby enabling the design of artificial proteins with specific functions. The research results have been published in the journal Nature Methods, providing a new research tool for the fields of biotechnology and medicine.

Proteins are the "molecular machines" of cells, composed of multiple domains (subunits) responsible for key functions such as energy metabolism and genetic material synthesis. Inspired by the evolution of proteins in nature, Professor Niopek's team used an AI model to simulate domain recombination and developed the open-source tool ProDomino. This tool analyzes datasets from over 100,000 proteins to predict optimal domain combinations for constructing novel proteins with controllable functions.

The research team has applied ProDomino to practical cases, such as combining chemically sensitive domains with the CRISPR-Cas system to generate controllable gene editing tools. Project leader Dr. Jan Mathony stated: "These hybrid proteins can be precisely turned on or off, improving the safety of genome editing." The model has currently been validated on various proteins and is open for use by global researchers.

Professor Niopek pointed out: "ProDomino not only simplifies the preparation of artificial proteins but also improves the precision of design, opening new paths for biomedical applications." In the future, this technology may drive the development of customized treatment methods.