metaAMR-Plus

Introduction

metaAMR-Plus is a Nextflow pipeline for comprehensive analysis of long-read Nanopore metagenomic data. It detects antimicrobial resistance (AMR) genes, identifies plasmids and virulence factors, and performs taxonomic classification using multiple parallel tools and databases. Results are integrated into a single interactive HTML report.

The pipeline supports two main modes:

• Standard mode — full analysis with assembly, AMR detection, plasmid detection, and taxonomic profiling
• Target species mode — extract and analyse reads from specific target organisms (e.g. known pathogens of interest)

Pipeline overview

Input reads (Nanopore FASTQ)
    │
    ├── Quality control (FastQC)
    ├── Adapter trimming (Porechop_ABI)
    ├── Length filtering (Filtlong)
    └── Host removal (Minimap2)
         │
         ├── [WITH ASSEMBLY]
         │    ├── Assembly (Flye) → [Polishing (Racon) →] QC (QUAST)
         │    ├── AMR detection:        RGI · AMRFinderPlus · ResFinder
         │    ├── Virulence detection:  Abricate (VFDB)
         │    ├── AMR integration:      hAMRonization
         │    ├── Plasmid detection:    PlasmidFinder · PlasClass
         │    └── Taxonomic profiling:  Centrifuge · Kaiju · Krona
         │
         ├── [WITHOUT ASSEMBLY]
         │    ├── ResFinder (reads-based)
         │    └── Taxonomic profiling:  Centrifuge · Kaiju · Krona
         │
         └── [TARGET SPECIES MODE]
              ├── Classification (Centrifuge)
              ├── Read extraction (target species only)
              └── AMR detection (ResFinder)
                   │
                   ▼
         MultiQC + Interactive HTML report

Features

• Multi-tool AMR detection — RGI (CARD), AMRFinderPlus, and ResFinder run in parallel; RGI, AMRFinderPlus, and Abricate results are integrated via hAMRonization
• Virulence factor detection — Abricate (VFDB database); results are included in hAMRonization
• Plasmid detection — PlasmidFinder (replicon typing) + PlasClass (sequence composition)
• Dual taxonomic profiling — Centrifuge (k-mer) and Kaiju (protein-level) with Krona visualisation
• Target species mode — extract reads classified as specific organisms before AMR analysis
• Interactive HTML report — integrated summary with AMR heatmap, VF categorisation, plasmid-taxonomy cross-reference, contig search
• Adaptive reporting — report automatically adjusts based on which tools were run
• Flexible tool selection — core tools run by default; enable assembly and assembly-based tools (RGI, AMRFinderPlus, PlasmidFinder, PlasClass) explicitly via --perform_assembly and --run_* flags

Quick start

1. Install Nextflow (≥24.04.2)
2. Install Singularity or Docker
3. Prepare your samplesheet (see usage)
4. Run the pipeline:

Minimal run

nextflow run Leilanasd/metaAMR-Plus \
    -profile <singularity/docker>
    --outdir results_test

Default run (FastQC + ResFinder + Centrifuge + Krona + MultiQC + HTML report)

nextflow run Leilanasd/metaAMR-Plus \
    -profile <singularity/docker> \
    --input samplesheet.csv \
    --outdir results \
    --databases database.csv

Full run (all tools enabled)

nextflow run Leilanasd/metaAMR-Plus \
    -profile <singularity/docker> \
    --input samplesheet.csv \
    --outdir results \
    --databases database.csv \
    --hostremoval_reference /path/to/host.fa \
    --perform_trim \
    --perform_hostremoval \
    --perform_assembly \
    --perform_polish_assembly \
    --run_abricate \
    --run_plasmidfinder \
    --run_plasclass \
    --run_rgi \
    --run_amrfinderplus \
    --run_kaiju \
    --download_rgi_db \
    --download_amrfinderplus_db

If you have existing local databases, provide them in database.csv instead of using the download flags.

Target species mode

To run analysis restricted to specific organisms:

nextflow run /path/to/metaAMR-Plus \
    -profile <singularity/docker> \
    --input samplesheet.csv \
    --outdir results_kleb \
    --target_species "Klebsiella pneumoniae" \
    --databases database.csv \
    --download_resfinder_db

Multiple species can be specified as a comma-separated string:

--target_species "Klebsiella pneumoniae,Enterococcus faecium"

Samplesheet format

sample,reads
LH040,/path/to/LH040.fastq.gz
LH024,/path/to/LH024.fastq.gz
LH077,/path/to/LH077.fastq.gz
LH085,/path/to/LH085.fastq.gz

Important: The sample field must match the base name of the FASTQ file (without .fastq.gz) for consistent MultiQC naming.

Databases

metaAMR-Plus uses several reference databases. These can be downloaded automatically by the pipeline or provided as pre-built local copies.

AMR databases

Tool	Database	Auto-download flag	Notes
RGI	CARD	--download_rgi_db	Or provide via database.csv
AMRFinderPlus	NCBI AMRFinder	--download_amrfinderplus_db	Or provide via database.csv
ResFinder	ResFinder DB	--download_resfinder_db	Or provide via database.csv

To save downloaded databases for reuse in future runs:

--save_databases

To use existing local databases (skip download), provide them all via the database.csv :

tool,db_name,db_params,db_path
kaiju,kaiju_db,,/path/to/kaiju_db
centrifuge,centrifuge_db,,/path/to/centrifuge_db
rgi,rgi_db,,/path/to/rgi_db
resfinder,resfinder_db,,/path/to/resfinder_db
amrfinderplus,amrfinderplus_db,,/path/to/amrfinderplus_db

Pass it with --databases database.csv. No download flags are needed when paths are provided in the database.csv.

Taxonomy and profiling databases

Centrifuge and Kaiju databases are always provided via the database.csv (see above). These cannot be auto-downloaded and must be built or obtained separately before running the pipeline.

Built-in databases (no setup required)

Tool	Database	Notes
Abricate	VFDB	Bundled with Abricate container
PlasmidFinder	PlasmidFinder DB	Bundled with PlasmidFinder container
PlasClass	—	Model-based, no external DB needed

Output

The main outputs are in <outdir>/:

Directory	Contents
report/	Interactive HTML report
multiqc/	MultiQC quality report
assemblies/	Raw assembly FASTA (Flye)
polished_assemblies/	Polished assembly FASTA (Racon)
quast/	Assembly quality statistics
resfinder/	ResFinder AMR results
rgi/	RGI (CARD) results
amrfinderplus/	AMRFinderPlus results
abricate/	Abricate VF results
hamronization/	hAMRonization integrated summary
plasmidfinder/	PlasmidFinder replicon results
plasclass/	PlasClass contig classification
centrifuge/	Centrifuge taxonomic profiles
kaiju/	Kaiju taxonomic profiles
krona/	Krona interactive visualisations

See output documentation for full details.

Credits

metaAMR-Plus was originally written by Leila Nasirzadeh and Jyotirmoy Das.

Team

• Leila Nasirzadeh — Department of Biomedical and Clinical Sciences, Linköping University
• Jyotirmoy Das — Department of Biomedical and Clinical Sciences, Linköping University

We thank the following people for their contributions to the development of this pipeline:

• Malgorzata Lysaik
• Jenny Welander
• Haiko Schurz
• Ida Karlsson
• Olivia Andersson
• Samuel Lampa
• Mårten Lindqvist

Citation

If you use this pipeline, please cite:

metaAMR-Plus: a Nextflow pipeline for comprehensive AMR analysis of Nanopore metagenomic data. DOI: 10.5281/zenodo.15682600

Please also cite the individual tools used. A full list is available in CITATIONS.md.

Acknowledgements

This pipeline was developed at the Bioinformatics Core Facility, Faculty of Medicine and Health Sciences,Linköping University, and Clinical Genomics Linköping, SciLifeLab. We are also grateful to LiU-IT (DIGIT) for their support on Fraka (HPC) during development and testing of metaAMR-Plus.

Contributions and support

If you encounter issues or have suggestions, please open a GitHub issue.