Anaerobic digestion (AD) is a promising biological process that converts waste into sustainable energy. To fully exploit AD's capability, we need to deepen our knowledge of the microbiota involved in this complex bioprocess. High-throughput methodologies open new perspectives to investigate the AD process at the molecular level, supported by recent data integration methodologies to extract relevant information. In this study, we investigated the link between microbial activity and substrate degradation in a lab-scale anaerobic codigestion experiment, where digesters were fed with nine different mixtures of three cosubstrates (fish waste, sewage sludge, and grass). Samples were profiled using 16S rRNA sequencing and untargeted metabolomics. In this article, we propose a suite of multivariate tools to statistically integrate these data and identify coordinated patterns between groups of microbial and metabolic profiles specific of each cosubstrate. Five main groups of features were successfully evidenced, including cadaverine degradation found to be associated with the activity of microorganisms from the order Clostridiales and the genus Methanosarcina. This study highlights the potential of data integration toward a comprehensive understanding of AD microbiota.
Keywords: 16S rRNA sequencing; codigestion; data integration; metabolomic; methanisation; partial least-squares regression.