From mechanism to application: Decrypting light-regulated denitrifying microbiome through geometric deep learning.

Area of Science:

• Microbiology
• Biotechnology
• Computational Biology

Background:

• Microbiome regulation is vital for industrial biotechnology and nitrogen cycling.
• Meta-omics offers genetic insights but faces challenges in complex data decryption.
• Precise control of denitrifying microbiomes is essential for sustainable applications.

Purpose of the Study:

• To develop a novel framework for encrypting and regulating denitrification microbiomes.
• To leverage geometric deep learning and optogenetics for microbiome analysis and control.
• To enhance microbial functions like nitrate removal and protein production.

Main Methods:

• Integration of optogenetics with geometric deep learning to create a discover-model-learn-advance (DMLA) cycle.
• Application of graph neural networks (GNNs) for biological knowledge integration and coexpression gene panel identification.
• Utilizing GNNs to predict phenotypes, elucidate mechanisms, and advance biotechnologies.

Main Results:

• The DMLA cycle successfully decrypted and regulated denitrification microbiomes.
• Graph neural networks demonstrated superior performance in analyzing complex meta-omics data.
• Discovery of a wavelength-divergent secretion system and nitrate-superoxide coregulation.
• Achieved an 83.8% increase in extracellular protein production and a 99.9% enhancement in nitrate removal.

Conclusions:

• GNN-empowered optogenetic approaches offer powerful tools for regulating denitrification.
• The DMLA cycle accelerates mechanistic discovery in microbiomes for diverse applications.
• This study advances sustainable industrial biotechnology and ecological nitrogen cycling through microbiome engineering.