Creating microbiome-model harmony between metaproteomics data and the ADM1da for a two-step anaerobic digester

Abstract

The effective operation, planning, and optimization of renewable energy production in anaerobic digestion (AD) plants relies on advanced process models, such as the Anaerobic Digestion Model No. 1 (ADM1). This study applies an ADM1-based model (ADM1da) to simulate a two-step digester in an industrial setting. The data demonstrate that 2.6% of the methane is lost as a result of open hydrolysis. Conversely, the incorporation of a hydrolysis fermenter enhances methane production by an average of 2.5%. Although ADM1-like models are widely recognized for accurately representing anaerobic digestion processes, mechanistic insights into the microbiome involved have been limited by the absence of tools to analyze microbial composition and functionality at the time these models were developed. To overcome this limitation, we utilized a metaproteomics approach to assess the abundance and biomass-correlated activity of microbial groups as defined by the model, aiming to bridge the gap between microbial ecology and bioprocess engineering in AD systems. We also developed and evaluated a series of rules for associating particular microbial species with functional groups of the model. Our analysis demonstrates that while the model supports the presence of a stable microbiome composition in the main fermenter, it is difficult to capture the dynamic behavior observed in the hydrolysis fermenter. Furthermore, the actual AD microbiome displays a greater versatility than the model assumes, with microorganisms performing multiple functions rather than being restricted to single roles. In conclusion, this study identifies options for improving AD models and integrating comprehensive biological knowledge to further optimize the performance of anaerobic digesters. HighlightsO_LISimulations revealed 2.6 % methane volume loss attributed to open hydrolysis C_LIO_LIImplementation of a two-step process increased methane production by 2.5% C_LIO_LIIdentification of rules to map metaproteomics data to ADM1da C_LIO_LISimulations of ADM1da depict the dynamic in the main but not in the hydrolysis fermenter C_LIO_LIMicrobial species perform multiple functions not just one as assumed in the ADM1da C_LI