Metagenomics data analysis: from mixed-community reads to biological insight.

1. Overview: what is metagenomics?

Metagenomics analyzes DNA recovered directly from mixed microbial communities without culturing individual organisms. It is used in human microbiome studies, environmental microbiology, pathogen surveillance, biotechnology, agriculture, wastewater monitoring, food safety and industrial process monitoring.

2. Main types of metagenomics studies

3. Experimental design

Metagenomics studies should be designed around sampling, controls, sequencing depth and expected biomass. Bias can enter before sequencing, during DNA extraction, library preparation, database classification or statistical analysis.

Questions to answer early

• Is the project 16S/ITS amplicon, shotgun metagenomics, viromics or genome-resolved metagenomics?
• What environment is being sampled: gut, soil, water, skin, oral, respiratory, wastewater, food, bioreactor or another setting?
• Are samples high-biomass or low-biomass?
• Which controls are included: extraction blanks, PCR blanks, mock communities and positive controls?
• Which confounders must be recorded: batch, extraction kit, storage, diet, medication, age, geography, time point or sequencing run?
• Is the primary endpoint taxonomy, function, diversity, pathogen detection, AMR profiling or MAG reconstruction?
• What database versions and taxonomic naming conventions will be used?

4. Metadata and controls

Metagenomics interpretation depends heavily on metadata. Technical variables can strongly influence observed community composition.

sample_id	group	body_site	batch	control_type	fastq_1	fastq_2
S1	control	gut	A	none	S1_R1.fastq.gz	S1_R2.fastq.gz
S2	control	gut	A	none	S2_R1.fastq.gz	S2_R2.fastq.gz
S3	treated	gut	B	none	S3_R1.fastq.gz	S3_R2.fastq.gz
S4	treated	gut	B	none	S4_R1.fastq.gz	S4_R2.fastq.gz
Blank1	control_blank	NA	A	extraction_blank	Blank1_R1.fastq.gz	Blank1_R2.fastq.gz
Mock1	mock	NA	A	mock_community	Mock1_R1.fastq.gz	Mock1_R2.fastq.gz

5. Input files and reference resources

6. FASTQ quality control

FASTQ QC checks sequencing quality before downstream profiling. In metagenomics, read quality, adapter contamination, sample read depth and control samples are all important.

mkdir -p results/qc/fastqc results/qc/multiqc

fastqc data/fastq/*.fastq.gz \
  --outdir results/qc/fastqc \
  --threads 8

multiqc results/qc \
  --outdir results/qc/multiqc

• Check read quality, adapter content and per-sample read counts.
• Compare control samples against true biological samples.
• Review GC-content distribution, which may be multimodal in mixed communities.
• Check whether paired-end reads are suitable for merging in amplicon workflows.

7. Adapter, primer and quality trimming

Trimming strategies differ for amplicon and shotgun workflows. Amplicon analysis often requires primer removal, while shotgun analysis usually focuses on adapters and low-quality tails.

mkdir -p results/trimmed results/qc/fastp

fastp \
  --in1 data/fastq/S1_R1.fastq.gz \
  --in2 data/fastq/S1_R2.fastq.gz \
  --out1 results/trimmed/S1_R1.trimmed.fastq.gz \
  --out2 results/trimmed/S1_R2.trimmed.fastq.gz \
  --html results/qc/fastp/S1.fastp.html \
  --json results/qc/fastp/S1.fastp.json \
  --thread 8

cutadapt \
  -g FORWARD_PRIMER_SEQUENCE \
  -G REVERSE_PRIMER_SEQUENCE \
  -o results/trimmed/S1_R1.primertrim.fastq.gz \
  -p results/trimmed/S1_R2.primertrim.fastq.gz \
  data/fastq/S1_R1.fastq.gz \
  data/fastq/S1_R2.fastq.gz

8. Host read removal

Human, animal or plant-associated metagenomics often includes host DNA. Host filtering reduces privacy risk, storage burden and false microbial assignments.

mkdir -p results/host_filtered results/logs

bowtie2 \
  -x reference/host/bowtie2/host_genome \
  -1 results/trimmed/S1_R1.trimmed.fastq.gz \
  -2 results/trimmed/S1_R2.trimmed.fastq.gz \
  --very-sensitive \
  -p 16 \
  --un-conc-gz results/host_filtered/S1_microbiome_reads.fastq.gz \
  -S /dev/null \
  2> results/logs/S1.host_filtering.log

9. 16S, ITS and amplicon analysis

Amplicon sequencing targets marker genes or regions. The analysis usually produces ASV or OTU tables, taxonomy assignments and diversity metrics.

10. Shotgun metagenomics

Shotgun metagenomics sequences all DNA in the community. It supports taxonomic profiling, functional profiling, strain analysis, assembly, MAG recovery and pathogen or AMR screening.

11. Taxonomic profiling

Taxonomic profiling estimates which organisms are present and their relative abundance. Different tools use different database structures and classification strategies.

mkdir -p results/taxonomy/kraken2 results/taxonomy/bracken

kraken2 \
  --db databases/kraken2_standard \
  --paired results/host_filtered/S1_R1.fastq.gz results/host_filtered/S1_R2.fastq.gz \
  --threads 16 \
  --report results/taxonomy/kraken2/S1.report \
  --output results/taxonomy/kraken2/S1.kraken

bracken \
  -d databases/kraken2_standard \
  -i results/taxonomy/kraken2/S1.report \
  -o results/taxonomy/bracken/S1.species.tsv \
  -r 150 \
  -l S

12. Functional profiling

Functional profiling estimates genes, pathways, enzymes or biological functions encoded by the community. It can be read-based, assembly-based or gene-catalog-based.

13. Alpha and beta diversity

Diversity metrics summarize within-sample richness and between-sample community differences. They are commonly used in 16S and shotgun metagenomics.

14. Metagenome assembly

Assembly combines shotgun reads into contigs. It can reveal genes, operons, plasmids, phages and partial or complete microbial genomes.

mkdir -p results/assembly/S1

megahit \
  -1 results/host_filtered/S1_R1.fastq.gz \
  -2 results/host_filtered/S1_R2.fastq.gz \
  -o results/assembly/S1 \
  --num-cpu-threads 16

15. Metagenome-assembled genomes

MAG analysis bins contigs into draft genomes based on coverage, composition and co-abundance patterns. MAGs can support taxonomy, metabolic reconstruction and genome-resolved ecology.

16. Antimicrobial resistance and virulence analysis

Metagenomics can screen for antimicrobial resistance genes, virulence markers and mobile genetic elements. Interpretation requires caution because gene detection does not necessarily prove expression, viability or clinical resistance.

17. Contamination and low-biomass interpretation

Contamination can come from reagents, kits, laboratory environment, barcode leakage, sample carryover or database misclassification. It is especially important in low-biomass samples.

18. Statistical analysis and differential abundance

Metagenomics data are sparse, compositional and often strongly affected by covariates. Statistical testing should match the data type and study design.

19. Visualization

Good metagenomics visualization should show both community structure and data quality. Avoid overcrowded stacked bar plots with too many rare taxa.

20. Example shotgun metagenomics workflow

The following simplified workflow illustrates a common shotgun metagenomics route. Real projects should adapt parameters to environment, host, database, controls and research question.

# 1. Raw-read QC
fastqc data/fastq/*.fastq.gz --outdir results/qc/fastqc --threads 8

# 2. Trim reads
fastp \
  --in1 data/fastq/S1_R1.fastq.gz \
  --in2 data/fastq/S1_R2.fastq.gz \
  --out1 results/trimmed/S1_R1.fastq.gz \
  --out2 results/trimmed/S1_R2.fastq.gz \
  --html results/qc/fastp/S1.html \
  --json results/qc/fastp/S1.json \
  --thread 8

# 3. Remove host reads
bowtie2 -x reference/host/bowtie2/host_genome \
  -1 results/trimmed/S1_R1.fastq.gz \
  -2 results/trimmed/S1_R2.fastq.gz \
  --very-sensitive -p 16 \
  --un-conc-gz results/host_filtered/S1_microbiome.fastq.gz \
  -S /dev/null

# 4. Taxonomic profiling
kraken2 --db databases/kraken2_standard \
  --paired results/host_filtered/S1_microbiome.1.fastq.gz results/host_filtered/S1_microbiome.2.fastq.gz \
  --threads 16 \
  --report results/taxonomy/S1.kraken2.report \
  --output results/taxonomy/S1.kraken2.output

# 5. Optional assembly
megahit \
  -1 results/host_filtered/S1_microbiome.1.fastq.gz \
  -2 results/host_filtered/S1_microbiome.2.fastq.gz \
  -o results/assembly/S1 \
  --num-cpu-threads 16

# 6. Summarize QC
multiqc results --outdir results/qc/multiqc_final

21. Deliverables and reporting

• FASTQ QC, trimming and host-filtering reports.
• Final MultiQC report with sample-level metrics and outliers.
• Host-read fraction and microbial-read retention summary.
• Taxonomic abundance tables at phylum, genus, species or strain level where appropriate.
• Functional abundance tables for genes, pathways, enzymes, AMR or virulence markers when requested.
• Alpha and beta diversity tables and plots.
• Control-sample contamination assessment.
• Assembly statistics, contigs and MAG reports when assembly is performed.
• Statistical analysis tables with effect sizes, adjusted p-values and model notes.
• Methods section with database names, versions, parameters and limitations.

22. Metagenomics analysis cheat sheet

Useful metagenomics tools and documentation

Frequently asked questions