Chapter 12: MAG Quality Assessment (CheckM) & Taxonomic Placement (GTDB-Tk)
Chapter 12: MAG Quality Assessment (CheckM) & Taxonomic Placement (GTDB-Tk)
A MAG may be a fragmented partial genome or a chimeric mixture of contigs from two distinct organisms. Before biological inference or public database submission, MAGs must be evaluated against strict quality benchmarks established by the Genomic Standards Consortium (MIMAG standards).
MIMAG Quality Tiers (Bowers et al.)
- High-Quality Draft MAG: >90% Completeness, <5% Contamination, presence of 5S, 16S, 23S rRNA genes and at least 18 tRNAs.
- Medium-Quality Draft MAG: ≥50% Completeness, <10% Contamination.
- Low-Quality Draft MAG: <50% Completeness or >10% Contamination (Must be discarded).
1. Assessing Completeness & Contamination with CheckM2
CheckM2 uses machine learning gradient-boosted decision trees to assess lineage-specific single-copy marker gene redundancy without taxonomic assignment priors.
# Run CheckM2 quality prediction across all generated bins
checkm2 predict \\
--input ./bins/ \\
--output-directory ./checkm2_out/ \\
--threads 16 \\
-x fa2. Phylogenomic Taxonomy with GTDB-Tk
GTDB-Tk places MAGs into the Genome Taxonomy Database reference tree using 120 bacterial or 53 archaeal marker proteins, providing standardized evolutionary taxonomy.
# Classify high-quality MAGs with GTDB-Tk
gtdbtk classify_wf \\
--genome_dir ./high_quality_bins/ \\
--out_dir ./gtdbtk_out/ \\
--cpus 16 \\
--extension fa