🦠 Reconstructing Uncultured Genomes from Environmental DNA
Shotgun metagenomics sequences all microbial DNA directly from soil, marine, gut, or host environments. Modern bioinformatic pipelines assemble contigs and group them into Metagenome-Assembled Genomes (MAGs) using tetranucleotide frequencies and differential co-abundance coverage profiles across samples.
🔬 CheckM2 MAG Quality Tier & MIMAG Standards
Interactive ToolSingle-Copy Marker Metrics
Calculated MIMAG Quality Tier
💾 Assembler RAM & Multi-k-mer Optimization
Resource AdvisorDataset Specs
Resource & Tool Recommendation
0. Shotgun Metagenomics vs 16S Marker Gene Sequencing
While 16S amplicon sequencing surveys only the bacterial hypervariable region, whole-metagenome shotgun (WMS) sequencing untangles full viral, bacterial, archaeal, and eukaryotic genomes, resolving single-nucleotide strain diversity and complete biosynthetic gene clusters (BGCs).
1. Quality Control & Host Decontamination
Clean raw paired-end reads and purge host contamination (human, mouse, or plant DNA) using Bowtie2.
# Adapter trimming and low-quality filtering fastp -w 8 -i raw_R1.fq.gz -I raw_R2.fq.gz \ -o clean_R1.fq.gz -O clean_R2.fq.gz --detect_adapter_for_pe # Remove human host reads with Bowtie 2 bowtie2 -p 8 -x hg38_index/genome \ -1 clean_R1.fq.gz -2 clean_R2.fq.gz \ --un-conc-gz microbial_R%.fq.gz -S /dev/null
2. De Novo Assembly with MEGAHIT
Assemble contiguous sequences (contigs) from millions of short reads using a succinct de Bruijn graph across multiple increasing k-mer sizes.
megahit -t 16 -m 0.85 \ -1 microbial_R1.fq.gz -2 microbial_R2.fq.gz \ --k-list 21,29,39,59,79,99,119,141 \ --min-contig-len 1000 \ -o megahit_assembly/
3. Assembly Quality Assessment with MetaQUAST
Evaluate assembly metrics: total length, contig count, largest contig, and N50.
metaquast.py megahit_assembly/final.contigs.fa \ -t 8 -o quast_report/
4. Contig Coverage Profiling (jgi_summarize_bam_contig_depths)
Align clean reads back to the assembled contigs to calculate base-level depth across samples. Differential co-abundance is essential for MetaBAT2 binning.
# Step 1: Index contigs and align reads bowtie2-build megahit_assembly/final.contigs.fa contigs_bt2 bowtie2 -p 16 -x contigs_bt2 -1 microbial_R1.fq.gz -2 microbial_R2.fq.gz | \ samtools sort -@ 8 -m 2G -o contig_coverage.sorted.bam - samtools index contig_coverage.sorted.bam # Step 2: Compute depth matrix jgi_summarize_bam_contig_depths \ --outputDepth depth.txt contig_coverage.sorted.bam
5. Metagenomic Binning with MetaBAT 2
Cluster contigs into individual putative organismal bins based on tetranucleotide frequencies (TNF) and coverage depth profiles.
metabat2 -t 8 \ -i megahit_assembly/final.contigs.fa \ -a depth.txt \ -m 1500 \ -o metabat_bins/bin
6. Bin Refinement & Dereplication (DAS Tool & dRep)
Combine multiple binners (MetaBAT2, MaxBin2, SemiBin2) into consensus non-redundant bins using DAS Tool, and dereplicate redundant genomes at 99% ANI with dRep.
# Dereplicate bins at species boundary (95% ANI) or strain boundary (99% ANI) dRep dereplicate drep_out/ \ -p 16 -g metabat_bins/*.fa \ -sa 0.95 -nc 0.30
7. CheckM2 Machine-Learning Quality Assessment
Evaluate MAG completeness and contamination using gradient-boosted decision trees trained on complete reference genomes.
checkm2 predict --threads 8 \ -x fa \ --input drep_out/dereplicated_genomes/ \ --output-directory checkm2_out/
8. GTDB-Tk Phylogenomic Taxonomic Classification
Assign standardized taxonomic lineages across 120 bacterial and 53 archaeal marker genes using FastTree and RED (Relative Evolutionary Divergence).
gtdbtk classify_wf \ --genome_dir drep_out/dereplicated_genomes/ \ -x fa \ --out_dir gtdbtk_out/ \ --cpus 16
9. Functional Annotation with Bakta & DRAM
Annotate CDS features, tRNA/rRNA genes, and reconstruct metabolic distillates (carbon, nitrogen, sulfur pathways).
# High-speed modern bacterial annotation with Bakta bakta --threads 8 --db bakta_db/ \ --output bakta_annotation/ \ --prefix MAG_01 \ drep_out/dereplicated_genomes/MAG_01.fa
10. Cohort MAG Abundance Profiling (CoverM)
Calculate relative abundance and trimmed mean coverage of each MAG across all sample cohorts.
coverm genome \ --genome-fasta-directory drep_out/dereplicated_genomes/ -x fa \ --bam-files sample_*.sorted.bam \ --methods relative_abundance trimmed_mean \ --threads 8 \ -o mag_abundance_table.tsv
🔀 Decision Architecture: Assembly & Binning Strategy
Decision TreeAssemble each sample individually to avoid chimeric contig collapse between distinct strains. Then, map reads from ALL samples across all assemblies to build a high-dimensional depth matrix for MetaBAT2.