Biomedical Data Analysis Platform

Home > Analysis Platform > Binning algorithm

Introduction：

An automated metagenome binning software tool to reconstruct single genomes from microbial communities for subsequent analyses of uncultivated microbial species. MetaBAT2 is a metagenome binning tool that groups contigs into putative genomes using tetranucleotide frequency and differential coverage across multiple samples. It employs an adaptive graph-based clustering algorithm that automatically determines optimal parameters, eliminating manual tuning. MetaBAT2 is fast, memory-efficient, and consistently ranks among top binners in benchmarks, especially when multi-sample coverage data is available.

Workflow Case Analysis Versions Manual Download

Input Loading: Reads contig sequences and per-sample depth matrix (from jgi_summarize_bam_contig_depths).
Feature Calculation: Computes TNF and normalized coverage profiles for each contig.
Graph Construction: Builds kNN graph where edges connect similar contigs based on combined TNF + coverage distance.
Adaptive Clustering: Iteratively partitions graph using label propagation; adjusts cluster granularity based on local density.
Output: Writes one FASTA file per bin + summary statistics.

Data

Results

Versions

Usage