Integrative analyses to investigate the link between microbial activity and metabolites degradation during anaerobic digestion

Abstract

Tel + 33 1 40 96 65 06 16 Fax + 33 1 40 96 61 99 17 olivier.chapleur@irstea.fr 18 https://orcid.org/0000-0001-9460-921X 19 20 Abstract 21Anaerobic digestion (AD) is a promising biological process to convert waste into sustainable 22 energy. However, the microbiota involved in this bioprocess is complex and additional 23 knowledge is still needed to fully exploit its capability. High throughput methodologies open 24 new perspectives, but innovative data integration methodologies are required for extracting 25 relevant information from these rich data. We analysed the association between microbial 26 activity and the patterns of substrate degradation during a lab-scale co-digestion experiment. 27These parameters were longitudinally monitored using 16S rRNA sequencing and untargeted 28 metabolomics. In this experiment, samples were collected from digesters fed with 9 different 29 mixtures of fish, sewage sludge, and grass. Our objective was to identify microorganisms 30 2 responsible for the degradation of molecules specific of each co-substrate. Five main groups 31 of correlated features were successfully evidenced. For example, the degradation of 32 cadaverine was found to be correlated with microorganisms from the order Clostridiales and 33 the genus Methanosarcina, and the degradation of lignin compounds was correlated with 34 cellulolytic degraders Lactobacillales. This study highlights the potential of data integration 35 towards a comprehensive understanding of AD microbiota. 36 37 Keywords 38 16S RNA sequencing; metabolomic; data integration; methanisation; co-digestion; PLS 39 canonical 40 41 101 4°C until the incubation experiments.102 Bioreactors experimental set-up 103 Binary mixtures of sludge with linearly increasing (0-100, 25-75, 50-50, 75-25, 100-0) 104 percentages of either fish or grass were prepared (Figure 1). Experiments were carried out in 105 51L glass bottles (700 mL working volume) at 35°C in the dark without agitation. The same 106 quantity of carbon was added in all the digesters, and the ratio of substrate/inoculum used to 107 feed and inoculate all the digesters was fixed at 12 gCOD/1.2 gCOD (Table S1). All the 108 bioreactors were complemented with a biochemical potential buffer (International Standard 109 ISO 11734 (1995)) to reach a final working volume of 700 mL. All incubations were 110 performed in triplicate. The bioreactors were then sealed with a screw cap and a rubber 111 septum and the headspaces were then flushed with N 2 (purity >99.99%, Linde gas SA). In 112 total 27 anaerobic bioreactors were set-up. 113Weekly (at days 0, 7, 14, 21), for every reactor, 6 mL of liquid phase were sampled 114 through the septum using a syringe. The collected samples were centrifuged at 10 000g for 10 115 minutes to separate the supernatants from the pellets. Supernatant were snap frozen using 116 liquid nitrogen and kept at -20°C for metabolomic analysis and pellets kept at -80°C for 117 microbial analysis. 118 119 RNA extraction and 16S rRNA sequencing 126Based on the biogas production, a t...

Not Just Numbers: Mathematical Modelling and Its Contribution to Anaerobic Digestion Processes

“…In data poor environments, where measurements are infrequent, erroneous or hard to obtain, mechanistic models are an obvious choice. Nevertheless, improvement in sensor technology and fast and cheap tools to characterise the microbial communities in digesters through multi-omic analyses, has gifted scientists and engineers with a wealth of data and knowledge of their ecology [108][109][110]. When used effectively, this can provide a greater resolution of understanding, helping to derive models that link microbial community composition and function effectively [111,112].…”

Not Just Numbers: Mathematical Modelling and its Contribution to Anaerobic Digestion Processes

“…In data poor environments, where measurements are infrequent, erroneous or hard to obtain, mechanistic models are an obvious choice. Nevertheless, improvement in sensor technology and fast and cheap tools to characterise the microbial communities in digesters through multi-omic analyses, has gifted scientists and engineers with a wealth of data and knowledge of their ecology [105][106][107]. When used effectively, this can provide a greater resolution of understanding, helping to derive models that link microbial community composition and function effectively [108,109].…”