Organic wastes (OWs) have increased 10-fold since the last century and are likely to double by 2025 as a result of the growth of the world population (Hoornweg et al., 2013). There is thus a strong need for sustainable OW management. From a waste management standpoint, aerobic (composting) and anaerobic digestions (AD) are the most obvious operational processes prior to soil applications. These recycling treatments can also be combined by performing a composting of digestates. A major constraint in these OW recycling scenarios is the capacity to assess the related OW health hazards, which are, in fact, dependent on their origin and contents in terms of chemical and microbiological contaminants, and the efficacy of the treatments at reducing these hazards.

AD implies a breakdown of organic materials naturally found in raw OW into biogas (50–75% CH₄, 25–50% CO₂, and 1–2% H₂S, H₂, and NH₃) and organic residues (called digestate) by a microbial consortium (Atelge et al., 2020). The success of this process relies on the synergies between microorganisms acting in a coordinated succession driven by the environmental changes occurring during the process (Rivière et al., 2009; Yu et al., 2010). To avoid deviation of microbial communities during the AD process, parameters such as volatile fatty acids (VFAs), ammonia and hydrogen sulfide concentrations, acetate:propionate ratios, and pH are regularly monitored (Kayhanian, 1994; Chen et al., 2008; González-Fernández and García-Encina, 2009; Wang et al., 2009; Franke-Whittle et al., 2014; Wagner et al., 2014). The microbial reactions of these processes can be performed by psychrophiles, mesophiles, and thermophiles (Moset et al., 2015; Saady and Massé, 2015; Hupfauf et al., 2018). Mesophilic digestion is considered the most stable AD process (Gou et al., 2014) because diversity of bacterial and archaeal assemblages are greater at mesophilic (35–37°C) temperatures (Levén et al., 2007; Pycke et al., 2011). The initial hydrolysis step of OW is performed by many facultative anaerobes or strictly anaerobic bacteria. This hydrolysis can degrade complex molecules into simple sugar monomers, amino acids, and alcohols. Fermentative bacteria will then convert these monomers into short-chain fatty acids, primary alcohols, as well as H₂ and CO₂. The production of acetate from the fatty acids will be performed by homoacetogenic bacteria like Butyribacterium and Acetobacterium, that of H₂ will be performed by the acetogenic microbiota such as Syntrophomonas, and that of CO₂ and H₂S will be performed by sulfate-reducing bacteria such as Desulfovibrio and Desulfobacter. Some of these compounds will then be used by methanogens to generate CH₄. Hydrogenotrophic methanogens such as Methanobacterium and Methanogenum sp. convert H₂ and CO₂ into CH₄, and acetotrophic ones such as Methanosarcina will convert acetate into CH₄ (Demirel and Scherer, 2008; Lykidis et al., 2011).

In the same way, composting is a natural process characterized by the conversion of OW into by-products that can be used as soil conditioners and/or organic fertilizers (Reinikainen and Herranen, 2001; Ahmad et al., 2007). This process is performed by an aerobic microbial consortium composed of Proteobacteria, Actinobacteria, and fungi. During the process of composting, microbial community successions are driven by environmental parameters such as moisture, temperature, and the aeration conditions (Blanc et al., 1999; Tian et al., 2013; Zhang et al., 2016). Composting in its initial phase can be carried out by several mesophiles including bacteria, which represent environmental health hazards such as E. coli, Salmonella, Klebsiella, and Nocardia (Insam and de Bertoldi, 2007). These organisms can break down soluble organic compounds by exothermic reactions. As the temperature rises above 40°C, these mesophiles will be outcompeted by more thermophilic bacteria such as Bacillus stearothermophilus and species from the Deinococcus/Thermus group (Insam and de Bertoldi, 2007). These will continue the breakdown processes and generate higher temperatures that will partly sanitize the end products by killing heat-sensitive hazardous microbial agents. A compost pile can reach 65–70°C. As the supply of high-energy compounds will get low, thermo-tolerant mesophiles will be favored and will perform the last transformation or “stabilization” reactions (Day and Shaw, 2001). The cool zones in the composting systems, particularly with those having no or little mixing, can favor a survival and (re)growth of some bacterial pathogens (Noble and Roberts, 2004).

Beyond the general microbial structures described previously for AD and composting, several questions remain, including how the high inputs of microorganisms and cells from the feeding substrates (OW) under conditions of continuous flow impact the microbial community structure in the end products. On one hand, some studies suggested that the microbial communities of digested OW are unique to a digester because of high differences in substrate and operating conditions (Werner et al., 2011; Sundberg et al., 2013; Zhang et al., 2014; De Vrieze et al., 2016). However, some of these analyses on full-scale AD plants fed with a mixture of substrates (De Vrieze et al., 2016) or with activated sludge (Wang et al., 2018) also indicated that the microbial communities could be stable over time. Nevertheless, in other studies, the microbial communities in AD were found variable over time (among a same digester), even when substrate and operational conditions remained constant (Supaphol et al., 2011; De Vrieze et al., 2013; Williams et al., 2013; Vanwonterghem et al., 2014). On the other hand, during composting, microbial communities of the raw OW were found to trigger a rise and decrease in temperature (Palaniveloo et al., 2020). These changes are highly selective but their incidence on microbial community patterns of mature compost remain to be evaluated according to the origin of OW. A knowledge gap thus remains about the relative contribution of the raw OW microbiome on the digestate and compost structural and functional microbial community patterns.

Here, we made the hypothesis that the microbiomes found among digesters and composts will be (i) significantly structured by the communities of the raw OW inputs, but that (ii) functional redundancy among these communities would generate distinct community profiles between treatment plants. Similar biodegradation activities between distinct treatment plants would thus be achieved by different microbial organizations because of this functional redundancy. To test these hypotheses, nine full-scale treatment plants located in various rural and urban backgrounds were investigated. These treatment plants were fed with agricultural or urban wastes or a mixture of both. Bacterial genetic structures were inferred from 16S rRNA gene meta-barcoding approaches, and patterns were compared between process types and according to the origin and nature of the raw OW. Functional traits were inferred from the 16S rRNA gene taxonomic allocations and databases recording the metabolic processes of well-defined bacterial taxa (FAPROTAX). The theoretical metabolic pathways of the genera impacted by treatment were further inferred from sequenced genomes using the MACADAM software (Le Boulch et al., 2019).

OW recycling of sewage sludge and livestock effluents represent a major challenge because not only the bioenergy production but also environmental and sanitary issues must be considered. A key criterion in the assessment of the microbiological risks of OW recycling that remain poorly defined is the nature and structure of the various microbiomes getting into these processes and their potential health and ecological hazards. Here, the core and flexible components of the end product microbiomes were inferred, and the likelihood of these at being reproducibly found among highly diverse OW treatment plants was evaluated. Nine full-scale treatment plants distributed over France were selected to obtain a global view of microbiome bacterial taxa in the French OW treatment sector. These inferred bacterial taxa (from metabarcoding 16S rRNA gene datasets) are discussed below according to process type (AD, composting) and OW origins (fecal wastes from WWTP or agricultural sites, and food products). Key functional traits of the bacterial genera impacted by the digestion processes were investigated to go further with the identification of key structuring bacterial processes.

According to NMDS ordinations of 16S rRNA gene metabarcoding datasets, two main bacterial community structures could be identified and found to match a segregation between bacterial taxa from OW containing agricultural wastes or not (restricted to urban WWTP activated sludge). This segregation was also in line with a differentiation between their contents in micropollutants particularly the TEs. The urban samples (treated or not) had significantly more micropollutants including 14 TEs (Ag, Al, As, Cd, Ce, Co, Cr, Cu, Ga, Mo, Ni, P, Pb, and Ti), 4-nonylphenol, and 2 PAHs (phenanthrene and anthracene) than the agricultural and mixed (agricultural/urban; termed “central” in the text) samples. The 16S rRNA gene microbial community structures of the investigated OW urban samples were in agreement with previous studies (Juretschko et al., 2002; Wong et al., 2005; Hu et al., 2012; Díaz et al., 2018). Inferred 16S rRNA gene taxonomic allocations were predominantly among the Proteobacteria, Bacteroidetes, Actinobacteria, Firmicutes, and Nitrospirae phyla. Proteobacteria were reported to be recurrent among WWTP activated sludge and have been shown to represent almost 50% of the bacterial taxa (Juretschko et al., 2002; Wong et al., 2005; Hu et al., 2012). Actinobacteria were also previously mentioned as highly abundant organisms in such wastes and being favored by the high phosphorus content of activated sludge (Wagner et al., 1994). Nitrospirae known to be nitrite-oxidizing bacteria were previously considered as common in activated sludge but were not mentioned as being abundant core organisms (Saunders et al., 2016).

After AD of both categories of raw OWs (urban and central/agricultural categories), 16S rRNA gene inferred phyla with similar relative changes were observed. These shifts occurred whatever the raw OW origin. Both categories of digestates (urban and central/agricultural categories) were impacted by selective environmental changes, which led to a decrease in the relative abundance of Proteobacteria but an enrichment in Firmicutes (which nearly doubled in relative abundance). Firmicutes had been reported by several authors as one of the most abundant phyla in AD processes of activated sludge (Hu et al., 2012; Zhang et al., 2012; Abendroth et al., 2015; Díaz et al., 2018). Sundberg et al. (2013) also found that Firmicutes were the most prevalent in samples taken from reactors fed with various combinations of wastes (from slaughterhouses, restaurants, households, etc.). In this work, with 52.2% of V5–V6 rRNA gene reads, Firmicutes were dominant in the digestates from central (mixed)/agricultural sites. Furthermore, they were enriched by 2.6-fold in digestates from urban sites to reach an 18.4% relative abundance. These results seem to indicate that the AD conditions favored their growth whatever the OW origins. To go deeper into the understanding of these community changes, the impact of AD on the structuration of end product microbiomes was assessed by grouping 16S rRNA gene reads according to their allocations at the genus level, and making comparisons by DESEQ2. These analyses revealed that about one third of the total relative abundance (according to the mean normalized counts) of the digestate bacterial taxa were significantly affected by the AD treatment whatever the raw OW origin. The relative abundances of 50 and 60 genera were significantly impacted by AD as observed from, respectively, urban and central/agricultural digestates. However, differences in the shifts of the 16S rRNA gene reads of these taxa were found related to the origin of the OW. In digestates of urban origin, this sorting mainly implied an increase in the relative abundance of OW urban taxa while those generated from agricultural/central OW led essentially to a reduction in the relative abundance of most of the initial OW taxa. This indicates that the coalescence of bacterial taxa of raw OW with the microbial community of the digesters will be largely dependent on their origin (and nature). In fact, urban OWs (mainly composed of activated sludge) were found to contain better fit communities for the AD process that could interact with the digesters’ microbial communities. Bacterial taxa present in agricultural raw OW were outcompeted by the digesters’ communities. This could be explained by the pre-treatment conditions performed in municipal wastewater facilities, leading to an increase of specialized bacterial taxa in the activated sludge. These taxa can perform multiple biodegradation processes (Shchegolkova et al., 2016) and can form heterogeneous structures termed flocs (Ju and Zhang, 2015). Functional inferences through FAPROTAX (Louca et al., 2016) were undertaken to highlight their potential contributions to the AD process.

The functional traits of the dominant microbial taxa inferred from the 16S rRNA gene datasets in the AD end products were investigated. Based on FAPROTAX analyses, functional traits such as chemoheterotrophy and fermentation were highly common among the taxa found in digestates of the urban and agricultural/central categories. Chemoheterotrophic bacteria are usually acting as primary decomposers (hydrolysis step), which are responsible for OW recycling in diverse ecosystems (Kämpfer et al., 1993). Fermentation, a widespread anaerobic pathway, is an essential step in AD process where fermentative bacteria convert simple monomers such as simple sugar, amino acids, and alcohols into short-chain fatty acids, primary alcohols, as well as H₂ and CO₂ (Cardinali-Rezende et al., 2009; Bajpai, 2017; Anukam et al., 2019). Interestingly, these functions (chemoheterotrophy and fermentation) were also previously reported by Wei et al. (2018) as the most dominant functions in activated sludge. This elucidates why bacterial genera from activated sludge could directly contribute to the observed diversity in the digestates. This observation is in line with the hypothesis of a selection of the most fit K-strategists (specialists) with ecologically relevant functions during the AD process. Also, these results confirmed that the changes in the AD community composition were driven by functional traits and less constrained by taxonomic relatedness.

To corroborate the above conclusion, closer analyses of the metabolic potential of bacterial communities were performed. Pathways inferred to be present in the enriched genera after treatment against those of the decreased ones were compared according to the origin/natures of the OWs. Interestingly, AD treatment was confirmed to impact similarly numerous metabolic pathways in both categories of OWs (urban and central/agricultural categories). Ten biosynthesis pathways (10/12) involved in the production of secondary metabolites, cell structure, metabolic regulators, biosynthesis, etc., and nine degradation pathways (9/11) (e.g., aromatic compounds and sulfur-containing compounds degradation) were affected similarly in both digestates from urban and agricultural/central categories. Nevertheless, some categories of pathways were not impacted in the same way according to the OW origin. For example, in digestates from agricultural and central origin, nitrogen cycling processes (e.g., nitrate-reduction, ammonia-assimilation, denitrification, and nitrogen-fixation pathways) were more abundant in the enriched genera than in the decaying ones, but the opposite situation was observed for digestates of urban origin. These pathways (related to nitrogen cycle) have been reported by Wei et al. (2018) as central key functions in activated sludge. Thus, these pathways could be critical in the ecology of activated sludge but less important during AD.

The impact of composting per OW type on the bacterial communities was also investigated. After composting, a NMDS plot grouped together the 16S rRNA community profiles whatever the OW origin. Composting was found to drive bacterial communities toward similar genetic structures in all samples without differentiating OW inputs that had previously undergone AD. This conserved microbiome appeared to have been selected by highly selective constraints leading to a significant decrease in the relative abundance of Bacteroidetes, Chlorobi, Chloroflexi, Lentisphaerae, Planctomycetes, Proteobacteria, Verrucomicrobia, and Spirochaetae and the loss of six phyla (Nitrospirae, Elusimicrobia, Acidobacteria, Saccharibacteria, TM6, and GOUTA4) from the inputs OW. Other studies reported similar findings (Steger et al., 2007a; Song et al., 2014). In a compost pile, changes in physicochemical factors such as pH, temperature, moisture, and organic carbon were found to be correlated with bacterial growth. Microbial community structures were found to be affected by these evolving parameters (Cahyani et al., 2003; Liang et al., 2003; Tang et al., 2007; Zhang et al., 2011). Among these, temperature was shown to be highly selective (Palmisano and Barlaz, 1996; Steger et al., 2007a). Interestingly, with their large shift in relative abundance (4.3-fold) observed between OW and composts, Actinobacteria showed a major re-shuffling with an up to 20% increase. These are clearly key players in the composting process.

To go deeper into the understanding of these effects, changes in the composition of the compost bacterial taxa were investigated at genus level. A total of 141 genera were found to be affected by composting, with 105 genera showing a significant decrease (equivalent to 53.7% of the total relative abundance of the compost’s bacterial taxa). Only three of these decaying genera were allocated to the Actinobacteria. However, 16 of the 36 enriched genera were part of this phylum. Previous studies have reported the persistence of Actinobacteria over the composting process from the thermophilic to the curing stage, regardless of the nature of treated OW (Danon et al., 2008). This bacterial group also constituted a significant part of the microbial community in the late composting stages (Herrmann and Shann, 1997; Tiquia et al., 2002; Hiraishi et al., 2003; Steger et al., 2007b). Actinobacteria are Gram-positive bacteria with high G + C contents in DNA, widely distributed in a diversity of ecological environments, including extreme ecosystem (Qin et al., 2011, 2019; Dhakal et al., 2017; Goodfellow et al., 2018). They can perform several important functions. Some Actinobacteria can synthesize a wide diversity of secondary metabolites, which may promote survival in highly selective environments (Fang et al., 2017; Puttaswamygowda et al., 2019; Qin et al., 2019). Several studies have highlighted the importance of temperature as a selective factor of thermo−tolerant Actinobacteria. In fact, Saccharomonospora viridis, Thermobifida fusca, and Thermobispora bispora have been reported in the later composting stages. In our study, many genera of thermo−tolerant Actinobacteria, e.g., Thermobifida, Actinomadura, Saccharomonospora, Arthrobacter, Thermomonoospora, and Thermocrispum were enriched in compost irrespective of the origin of the treated OW. Actinobacteria are capable of breaking down a wide range of animal wastes and plant debris particularly complex compounds (e.g., lignocelluloses), playing an important role in the later stages of composting (Herrmann and Shann, 1997; Wang et al., 2016). It is to be noted that thanks to the occurrences of bacterial genera degrading xenobiotics, composting offers the advantage of decreasing the concentration of chemical contaminants (Semple et al., 2001; Ghaly., 2011; Cascant et al., 2016; Ren et al., 2018). In line with these reports, genera of xenobiotic degrading Actinobacteria, e.g., Arthrobacter, were found to be enriched in the 16S rRNA gene datasets generated from composts DNA. To go deeper, closer analyses of functional potentials per taxa enriched after composting were performed. The functional profiles of the enriched genera suggested that the number of 16S rRNA copies of genera, e.g., Arthrobacter, Alcanivorax, Rhodococcus, Brevibacterium, etc., known to consist of many strains having functional traits involved in arsenate detoxification, dissimilatory arsenate reduction, plastic degradation, and oil bioremediation have significantly increased through composting. This further supports the idea that composting is an efficient bioremediation process.

In conclusion, AD and composting induced bacterial changes that were found driven by pH, temperature, but also the C- and energy sources which require specialized metabolic properties. These factors were more selective in composting than AD. Consequently, composting generated highly predictable community patterns whatever the OW origin. In AD, the relative contribution of the raw OW microbiomes on the digestate structural and functional bacterial community patterns was dependent on its origin. Bacterial genera from activated sludge were found to be well-suited for the AD process and contributed directly to the observed diversity observed in the digestates. Content in activated sludge among the OW was thus considered to be a critical factor for predicting the digestate bacterial contents. Our findings provided new insights into our understanding of bacterial coalescence phenomena in aerobic and anaerobic digesters. However, knowledge gap remains regarding potential health hazards associated with the emerging communities observed in the end products. These organic matter biodegradation processes being built from a large panel of exogenous communities, they might offer conditions favorable for horizontal gene transfer (HGT) events such as the HGT of antibiotic resistance genes (ARG). The next step in these analyses will be to evaluate the frequency of these HGT.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: European Nucleotide Archive (https://www.ebi.ac.uk/research) under the accession number PRJEB40193.

AA, EB, LM, BC, and WG conceived and designed experiments. AA, EB, LM, SH, DP, ED, BC, and WG performed the experiments. AA, EB, BC, and WG performed the data curation and statistical analyses. AA, BC, and WG wrote the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.