Background
So far, a lot of binning approaches have been intensively developed for untangling metagenome-assembled genomes (MAGs) and evaluated by two main strategies. The strategy by comparison to known genomes prevails over the other strategy by using single-copy genes. However, there is still no dataset with all known genomes for a real (not simulated) bacterial consortium yet.
Results
Here, we continue investigating the real bacterial consortium F1RT enriched and sequenced by us previously, considering the high possibility to unearth all MAGs, due to its low complexity. The improved F1RT metagenome reassembled by metaSPAdes here utilizes about 98.62% of reads, and a series of analyses for the remaining reads suggests that the possibility of containing other low-abundance organisms in F1RT is greatly low, demonstrating that almost all MAGs are successfully assembled. Then, 4 isolates are obtained and individually sequenced. Based on the 4 isolate genomes and the entire metagenome, an elaborate pipeline is then in-house developed to construct all F1RT MAGs. A series of assessments extensively prove the high reliability of the herein reconstruction. Next, our findings further show that this dataset harbors several properties challenging for binning and thus is suitable to compare advanced binning tools available now or benchmark novel binners. Using this dataset, 8 advanced binning algorithms are assessed, giving useful insights for developing novel approaches. In addition, compared with our previous study, two novel MAGs termed FC8 and FC9 are discovered here, and 7 MAGs are solidly unearthed for species without any available genomes.
Conclusion
To our knowledge, it is the first time to construct a dataset with almost all known MAGs for a not simulated consortium. We hope that this dataset will be used as a routine toolkit to complement mock datasets for evaluating binning methods to further facilitate binning and metagenomic studies in the future.