Why
A metagenome is DNA from a whole community of organisms, sequenced together. Binning groups the assembled fragments (contigs) into genomes. Dozens of binners exist, but they are hard to compare: each one bundles its own preprocessing, model and clustering, and runs on its own data.
What MegoBin does
- Benchmarks real pipelines. mmlong2 runs on the 20 MicroBench samples (a mock community, human gut, wastewater, sediment and soil) on AAU's BioCloud, measuring which binners win high-quality genomes and what each costs in CPU hours and memory.
- Takes methods apart. A re-implementation of SemiBin2's long-read, self-supervised mode, checked step by step against the original.
- Audits the field. A code-level review of 12 published binners against one shared interface.
Status
[Current milestone, e.g. v2: from fixed slots to a benchmarked DAG; link to repo once public.]