Background

Extremophiles are a remarkable group of microorganisms that thrive in environments previously considered inhospitable, such as extreme temperatures, salinity, and pH levels. These organisms hold immense potential for applications in biotechnology, environmental management, and synthetic biology due to their unique metabolic pathways and enzyme production. Extremozymes, enzymes produced by extremophiles, retain high catalytic efficiency under extreme conditions, making them invaluable in industries such as food processing, biofuel production, and industrial biocatalysis.

However, traditional methods of screening and identifying extremophiles are time-consuming and labor-intensive, relying heavily on manual cultivation and selection. To address this challenge, we developed a deep learning model called iExtreme that leverages genomic sequences to rapidly identify extremophiles and predict their optimal living conditions. Additionally, we curated a comprehensive dataset of 1,030 extremophilic genomes and employed machine learning techniques to discover novel extremozymes. Through phage-assisted non-continuous evolution (PANCE) technology, we further optimized these extremozymes, significantly enhancing their catalytic activity and stability.

This research provides a robust tool for the rapid identification of extremophiles and opens new pathways for the development and optimization of extremozymes, driving their potential applications in industrial and environmental sectors.