CheckM2
Metagenomic BinningCheckM2 is a rapid machine-learning-based tool for assessing the quality of metagenome-assembled genomes (MAGs). It uses gradient-boosted models trained on genomic features to predict completeness and contamination without requiring reference genomes. [1]
How to Obtain Output Model File
Below is a brief workflow the team ran to obtain the output model examples we present on the tools page.
Input
Directory of genome bins in FASTA format (.fa, .fna, .fasta)
Output
TSV file with per-bin quality metrics: completeness, contamination, coding density, GC content, genome size, N50, contig count
conda install -c bioconda checkm2
Docker image: nanozoo/checkm2:latest
Sample 1 21 metagenome-assembled bins from the PRJNA795985 gut microbiome co-assembly assessed with CheckM2 using the Database.
- 1
Download 3 samples from PRJNA795985
for SRR in SRR17531757 SRR17531762 SRR17531772; do fastq-dump --split-files --gzip -X 1000000 --outdir reads/ $SRR; done
1M read pairs per sample from "Diet and Antimicrobial Resistance in Healthy US Adults" (Shrestha et al., mBio 2022).
- 2
Co-assemble with MEGAHIT
megahit -1 reads/SRR17531757_1.fastq.gz,reads/SRR17531762_1.fastq.gz,reads/SRR17531772_1.fastq.gz -2 reads/SRR17531757_2.fastq.gz,reads/SRR17531762_2.fastq.gz,reads/SRR17531772_2.fastq.gz -o assembly --min-contig-len 1500 -t 8
Co-assembly produces 14,934 contigs (≥1,500 bp).
- 3
Map each sample to contigs
for SRR in SRR17531757 SRR17531762 SRR17531772; do minimap2 -ax sr -t 8 assembly/final.contigs.fa reads/${SRR}_1.fastq.gz reads/${SRR}_2.fastq.gz | samtools sort -@ 4 -o bams/${SRR}.sorted.bam && samtools index bams/${SRR}.sorted.bam; doneProduces per-sample BAM alignments required for binning.
- 4
Bin contigs
vamb bin default --outdir vamb_out --fasta assembly/final.contigs.fa --bamdir bams/ -m 1500 --minfasta 200000 -p 8
Produces 21 bin FASTA files (≥200 Kbp).
- 5
Download CheckM2 database
checkm2 database --download --path checkm2_db/
Downloads the DIAMOND database (uniref100.KO.1.dmnd, ~1.74 GB).
- 6
Run CheckM2 on all 21 bins
checkm2 predict --input bins/ --output-directory checkm2_output/ -x fna --threads 8 --database_path checkm2_db/uniref100.KO.1.dmnd
Produces quality_report.tsv with completeness, contamination, coding density, GC content, N50, and genome size for all 21 bins.
quality_report.tsv to IntMetaMaterials Used
Sample Output Files
Download the output files used in the tool page demos. You can upload these directly to IntMeta to explore the visualizations.
Single & Comparison
Group Analysis
Charts Reference
Detailed descriptions for all 32 visualizations generated by CheckM2 in IntMeta.
completeness-vs-contamination
Scatter plot of ML-predicted completeness (%) vs contamination (%) for each bin, colored by MIMAG quality tier. Ideal bins sit in the top-left corner. Dashed lines mark MIMAG thresholds: High quality (>90% complete, <5% contamination) and Medium quality (≥50% complete, <10% contamination).
quality-score-ranking
Horizontal bar chart ranking bins by composite quality score = Completeness − 5 × Contamination (the standard MIMAG formula). A bin with 95% completeness and 2% contamination scores 85. Color reflects the MIMAG quality tier (High / Medium / Low).
quality-metrics-heatmap
Heatmap of min-max normalized values of Completeness, Contamination, Coding_Density, and GC_Content across all bins. Each metric is scaled 0–1 within its column; darker cells indicate higher relative values. Useful for spotting outlier bins at a glance.
genome-size-distribution
Bar chart of Genome_Size (Mbp) per bin, sorted by size and colored by MIMAG quality tier. Dashed reference lines indicate typical bacterial genome size ranges (most are 1–8 Mbp). Unusually small or large bins may warrant manual inspection.
gc-content-vs-genome-size
Scatter plot of GC_Content (%) vs Genome_Size. Different taxa cluster at characteristic GC/size combinations — e.g., Firmicutes tend toward 30–40% GC while Actinobacteria often exceed 60%. Outlier positions may indicate chimeric bins mixing contigs from unrelated organisms.
n50-vs-genome-size
Bubble plot where X = Genome_Size, Y = Contig_N50 (half the assembly length is in contigs ≥ this size), and bubble diameter = Completeness (%). Top-right, large bubbles represent well-assembled, complete genomes.
coding-density-vs-completeness
Scatter plot of Coding_Density (fraction of bases in coding sequences) vs Completeness, colored by quality tier. Most prokaryotic genomes have coding density between 85–95%. Low coding density combined with high completeness may signal contamination with non-coding DNA.
assembly-fragmentation
Scatter plot of Total_Contigs (contig count) vs Genome_Size. Highly fragmented assemblies appear upper-left (many contigs, small genome). Well-assembled genomes cluster lower-right (few contigs, large genome). Fragmentation affects downstream analyses like gene prediction.
quality-tier-distribution
Bar chart counting bins in each MIMAG quality tier: High (>90% completeness, <5% contamination), Medium (≥50% completeness, <10% contamination), Low (below both thresholds). The most common summary chart in MAG publications.
completeness-threshold-recovery
Grouped bar chart showing how many bins pass ≥50% / ≥70% / ≥90% completeness at <5% and <10% contamination limits. Standard benchmarking visualization from CAMI challenges, DAS Tool, SemiBin2, and VAMB papers.
gc-content-histogram
Histogram of GC content in 5% bins. Bimodal or multimodal distributions suggest recovery of MAGs from distinct taxonomic groups, since different taxa tend to cluster at characteristic GC ranges.
cumulative-genome-recovery
Cumulative step-area plot with bins sorted by quality score (Completeness − 5×Contamination, descending). Y-axis shows running total of genome size in bp. A steep initial rise indicates most assembled sequence comes from high-quality bins.
metric-correlation-heatmap
Symmetric Pearson correlation matrix across all numeric quality metrics. Red = positive correlation, blue = negative. Reveals which metrics co-vary (e.g., completeness vs genome size) and which are independent, useful for understanding MAG quality drivers.
comp-quality-tiers
Grouped bar chart comparing the distribution of MIMAG quality tiers (High / Medium / Low) across samples. Each group shows one bar per sample, enabling quick assessment of which sample produced higher-quality bins overall.
comp-genome-size
Box plot or grouped bar chart comparing the genome size distribution of bins across samples. Reveals whether certain samples yield systematically larger or smaller MAGs, which may reflect community composition differences.
comp-completeness
Box plot comparing the distribution of ML-predicted completeness (%) across samples. Higher medians and tighter interquartile ranges indicate more consistently complete bins in that sample.
comp-contamination
Box plot comparing contamination (%) distribution across samples. Lower values are better — samples with elevated contamination may need stricter binning parameters or additional refinement.
comp-summary-table
Aggregate summary table showing key metrics per sample: total number of bins, mean completeness, mean contamination, mean genome size, and number of high-quality bins. Provides a quick numerical overview for cross-sample comparison.
comp-quality-pct
100% stacked bar chart showing the proportion of High / Medium / Low quality bins per sample. Normalizes for different bin counts, enabling direct comparison of MAG recovery quality. Standard in CAMI challenge benchmarks.
comp-cdf
Empirical cumulative distribution function (CDF) of bin sizes overlaid per sample. Reveals full distribution shape beyond boxplot summaries — tails, bimodality, and what fraction of bins exceed size thresholds. Standard in CAMI benchmarking.
comp-gc-content
Boxplot of GC content (%) distribution per sample. GC profiles differ across environments (soil vs. gut vs. marine) and extraction protocols, serving as a proxy for community composition differences.
comp-contig-n50
Boxplot of contig N50 per bin across samples. N50 is the primary assembly contiguity metric — higher values indicate less fragmented bins. Reported in virtually every MAG quality assessment.
comp-coding-density
Boxplot of coding density (fraction of genome encoding proteins) per sample. Lower coding density can indicate contamination with non-coding DNA or eukaryotic sequences. Evaluated by CheckM2 ML models.
binning-group-quality-tiers
Grouped bar chart of MIMAG quality tier counts (High / Medium / Low) per experimental group, with Chi-square test for independence. Tiers follow Bowers et al.: High (>90% completeness, <5% contamination), Medium (≥50%, <10%). Significant Chi-square p-values indicate that MAG quality distributions differ between groups.
binning-group-metric-boxplots
Boxplots comparing a selected quality metric (Completeness, Contamination, Genome Size, GC Content, N50, or Coding Density) across groups. All individual bin values are pooled per group. Kruskal-Wallis test (≥3 groups) or Mann-Whitney U test (2 groups) assesses significance; pairwise comparisons use Benjamini-Hochberg FDR correction.
binning-group-recovery-rate
Grouped bar chart of mean bin count per sample at each quality tier (High, Medium, Low, Total) across groups, with error bars showing ±1 SD. Tests whether groups differ in how many MAGs they recover — a key benchmark in the CAMI challenge framework.
binning-group-pca
PCA ordination of per-sample quality profiles (mean completeness, contamination, genome size, GC, N50, coding density, %HQ, %MQ). Features are z-score standardized before decomposition. PERMANOVA on the Euclidean distance matrix tests whether group centroids differ significantly. 95% confidence ellipses are drawn for groups with ≥3 samples.
binning-group-cdf
Overlaid empirical cumulative distribution functions (CDFs) of a selected metric across groups. The two-sample Kolmogorov-Smirnov test quantifies the maximum vertical distance between curves — significant D-statistics indicate the groups' metric distributions differ in location, spread, or shape.
binning-group-scatter
Cross-group scatter overlay of quality metrics, colored by group. Background zones mark quality tiers. Shows whether experimental groups produce systematically different MAG quality profiles. Based on MIMAG standards.
binning-group-total-recovery
Stacked bar chart of mean total base pairs recovered per group, split by quality tier. Error bars show ±1 SD. Measures actual data volume binned per group, complementing bin count and recovery rate charts.