Possibilities and obstacles in recovery of genomes from elusive microbes in complex metagenomes

Representative genomes provide an entry point for understanding a given ecosystem. The genomes themselves give insights in the metabolic potential and possible role of the bacteria in the ecosystem, as well as being essential when applying other omics based techniques. Metagenomics and single cell genomics can be used to access genomes from uncultured bacteria in complex environments, but both strategies have limitations when the complexity is high and micro-diversity is present. In this study, we explored the potentials and obstacles faced when assembling genomes from complex metagenomes using activated sludge as a model system. Activated sludge is a great model system for studying ecological principles, but despite its commonplace, little is known about the majority of the community members as they elude cultivation. A combination of short-term laboratory enrichments and sequence independent metagenomic binning was used to reduce community complexity and better target unknown species.
As a brute force strategy to enrich bacteria from different functional groups and reduce micro-diversity, activated sludge from a municipal wastewater treatment plant was incubated in batch cultures under six different growth conditions (O2 /-O2 x Glucose/Butyrate/Casein) for 7 days at 20 °C. Interesting samples were identified using Illumina V4 16S amplicon time series of each growth condition. Twelve enrichment samples were subsequently selected for metagenomics and sequenced on the Illumina platform (300 Gbp data in total). Metagenomes with reduced complexity, containing interesting species were used for de novo metagenome assembly in CLC Genomics Workbench v6.5. The metagenome scaffolds were binned to individual genomes using the multi-metagenome approach and interesting genomes extracted. To support the observations from the metagenomes we simulated different scenarios of micro-diversity and explored the impact on de novo assembly.
The short-term enrichments decreased the overall complexity and changed the relative abundances of the different community members, which improved the multi-metagenome binning. This made it possible to extract high quality genome bins from a range of different taxonomic groups, all abundant in the activated sludge systems (i.e. novel species in genus Ferruginibacter and the family Acidimicrobiaceae). There was widespread micro-diversity especially among the most high abundant community members, which resulted in highly fragmented genome assemblies and poor binning. However, the strains that actively grew during enrichment generally assembled better, due to relatively less micro-diversity. In silico analysis indicated that strain similarity of 96-99% produced critically fragmented assemblies for the assembly conditions used.
In conclusion, we showed that it was impossible to recover high quality genomes from specific species in complex samples with significant micro-diversity and that great care should be taken when analyzing assembled data from complex metagenomes. However, the use of short-term enrichments can be a successful way to lower the micro-diversity and hence greatly improve genome recovery from otherwise elusive microbes.