The anaerobic degradation of complex organic substrates involves a network of coupled steps catalyzed by Bacteria and Archaea. Complete degradation to methane and carbon dioxide largely relies on the structure and composition of the microbial communities. The complete anaerobic foodweb is known in the engineering world under the term anaerobic digestion and is currently facing renewed attention in the field of bioenergy production. Anaerobic digesters are most frequently empirically inoculated with a microbial community that has proven to produce methane under the desired operating conditions. Whether this microbial community compared to communities from other natural or engineered ecosystems produces the highest methane yield or possesses otherwise more beneficial properties has rarely been demonstrated. It may be that a more educated selection of the inocula for anaerobic digesters can further improve process performance. In our experiment, we therefore screen the metabolic potential of a wide range of inocula in a systematic and standardized procedure. All inocula are incubated with the same radiation-sterilized complex substrate containing cellulosic fibers and polysaccharides (20%), proteins (15%) and lipids (3.5%) suspended in a phosphate buffer at pH 7.5. We adapt the initial community to the complex substrate in semi-continuous reactors over a period of at least 60 days. Several pulses of substrate at a sufficiently high substrate to microorganism ratio of 3 g substrate/g living biomass allow the development of an active adapted community with stable performance towards the end of the incubations. The comparison of the initial and adapted communities by high-throughput sequencing will reveal on the molecular level the changes in the community during the adaptation phase. The adapted community structure is then related to ecosystem function (i.e. methane yield and production kinetics, volatile fatty acid composition) and serves as criterion to classify the microbial communities between highly productive and non-performing ecosystems. By correlating properties of the microbial communities (e.g. distributions of rank-abundance, apparent taxonomic richness/diversity, phylogenetic diversity, presence/abundance of core species) with the performance of ecosystem functions, we attempt to identify generally applicable molecular indicators that will be useful in the assessment of communities in managed microbial ecosystems but also when evaluating stability of ecosystem function in natural systems. This project is funded as a small project by the INRA metaprogram MEM.