Temporal Dynamics of Microbial Communities in Anaerobic Digestion: Influence of Temperature and Feedstock Composition on Reactor Performance and Stability

TL;DR

This study explores how temperature and feedstock composition influence microbial communities and performance in anaerobic digestion, using detailed chemical fingerprinting, microbiome analysis, and machine learning to improve understanding and predictability of reactor stability.

Contribution

It introduces a novel integrated approach combining chemical fingerprinting, microbiome sequencing, and machine learning to analyze and predict anaerobic digestion performance.

Findings

Mesophilic conditions support more diverse microbiomes.

High butyric acid correlates with increased Methanomassiliicoccus and reduced biogas.

Operational parameters and chemical composition significantly influence reactor performance.

Abstract

Anaerobic digestion (AD) offers a sustainable biotechnology to recover resources from carbon-rich wastewater, such as food-processing wastewater. Despite crude wastewater characterisation, the impact of detailed chemical fingerprinting on AD remains underexplored. This study investigated the influence of fermentation-wastewater composition and operational parameters on AD over time to identify critical factors influencing reactor biodiversity and performance. Eighteen reactors were operated under various operational conditions using mycoprotein fermentation wastewater. Detailed chemical analysis fingerprinted the molecules in the fermentation wastewater throughout AD including sugars, sugar alcohols and volatile fatty acids (VFAs). Sequencing revealed distinct microbiome profiles linked to temperature and reactor configuration, with mesophilic conditions supporting a more diverse and densely connected microbiome. Significant elevations in Methanomassiliicoccus were correlated to high butyric acid concentrations and decreased biogas production, further elucidating the role of this newly discovered methanogen. Dissimilarity analysis demonstrated the importance of individual molecules on microbiome diversity, highlighting the need for detailed chemical fingerprinting in AD studies of microbial trends. Machine learning (ML) models predicting reactor performance achieved high accuracy based on operational parameters and microbial taxonomy. Operational parameters had the most substantial influence on chemical oxygen demand removal, whilst Oscillibacter and two Clostridium sp. were highlighted as key factors in biogas production. By integrating detailed chemical and biological fingerprinting with ML models this research presents a novel approach to advance our understanding of AD microbial ecology, offering insights for industrial applications of sustainable waste-to-energy systems.