GraphBin

GraphBin refines metagenomic contig binning by leveraging assembly graph connectivity to improve bin assignment accuracy for microbial community species analysis.


Key Features:

  • Assembly graph utilization: Leverages assembly graph connectivity information produced during metagenomic assembly to inform binning decisions.
  • Supported graph types: Operates on underlying assembly graphs such as de Bruijn graphs or overlap-layout-consensus graphs.
  • Label propagation algorithm: Applies a label propagation algorithm on the assembly graph to propagate bin labels among connected contigs.
  • Refinement of external bins: Accepts initial binning results from other established binning tools and refines those assignments using graph connectivity.
  • Correction of mis-binned contigs: Identifies and corrects contigs that were assigned to incorrect bins in initial binning outputs.
  • Recovery of short contigs: Predicts labels for previously discarded short-length contigs based on graph connections.
  • Contrast to traditional signals: Uses graph connectivity rather than relying solely on oligonucleotide/k-mer composition and contig coverage for refinement.

Scientific Applications:

  • Metagenomic bin refinement: Improves bin assignment accuracy in metagenomic assemblies to produce more accurate MAGs (metagenome-assembled genomes).
  • Error correction: Detects and corrects mis-binned contigs from initial binning tools to improve downstream taxonomic and functional analyses.
  • Short-contig labeling: Recovers labels for short-length contigs that are often excluded, increasing sequence retention for community analyses.
  • Microbial community analysis: Enhances species-level classification and diversity inference in microbial ecology studies by refining contig bins.

Methodology:

GraphBin takes initial binning results, uses assembly graphs (de Bruijn or overlap-layout-consensus) generated during metagenomic assembly, and applies a label propagation algorithm over the graph to reassign or assign bin labels, correcting mis-binned contigs and labeling short-length contigs.

Topics

Details

License:
GPL-2.0
Tool Type:
command-line tool
Programming Languages:
Python
Added:
1/18/2021
Last Updated:
1/25/2021

Operations

Publications

Mallawaarachchi V, Wickramarachchi A, Lin Y. GraphBin: refined binning of metagenomic contigs using assembly graphs. Bioinformatics. 2020;36(11):3307-3313. doi:10.1093/bioinformatics/btaa180. PMID:32167528.