ARDETYPE

ARDETYPE is a modular, scalable pipeline for bacterial whole-genome sequencing (WGS) analysis. It supports Illumina (paired-end) and Oxford Nanopore data in fastq.gz format, as well as PacBio reads and pre-assembled genomes in FASTA format.
Designed for clinical microbiology and public health applications, it enables streamlined and reproducible analysis of bacterial isolates through both species-agnostic and species-specific workflows.

---

Features

• Species-agnostic and species-specific analysis of bacterial WGS data
• Modular, Snakemake-based workflow
• Automated quality control, assembly, classification, AMR profiling, plasmid typing, and reporting

---

Pipeline Modules

ARDETYPE consists of four main modules:

Module	Description
bact_core	Core tasks: QC, contamination filtering, species ID, assembly
bact_shell	General tools: resistance gene & plasmid detection, assembly QC
bact_tip	Species-specific typing and subtyping
bact_shape	Aggregation, summarization, and report generation

---

Supported Tools

Tool	Module
agrvate	bact_tip
amrfinder+	bact_shell
capybara	bact_tip
chewbbaca	bact_tip
circlator	bact_core
ectyper	bact_tip
emmtyper	bact_tip
fastp	bact_core
filtlong	bact_core
flye	bact_core
hicap	bact_tip
kleborate	bact_tip
kraken2	bact_core
legsta	bact_tip
lissero	bact_tip
lrefinder	bact_tip
medaka	bact_core
meningotype	bact_tip
mlst	bact_shell
mob-suite	bact_shell
plasmidfinder	bact_shell
polca	bact_core
polypolish	bact_core
quast	bact_shell
resfinder	bact_shell
rgi	bact_shell
seqsero2	bact_tip
seroba	bact_tip
shigatyper	bact_tip
shovill	bact_core
sistr	bact_tip
snikt	bact_core
spatyper	bact_tip
staphopia-sccmec	bact_tip
stecfinder	bact_tip
virulencefinder	bact_shell
hybracter	bact_core
cgmlst-dists	bact_shape
AcinetobacterPlasmidTyping	bact_tip
Hybracter	bact_core
rmlst API	bact_shell
PLSDB API	bact_shell

---

Installation

git clone -b standalone_rework https://github.com/NMRL/Ardetype
cd Ardetype
conda env create -f ardetype.yaml

• Note: Currently PLSDB API module requires some adjustments to the source code in order to make it work.
• Finding python module files in conda environment.

Download pre-packaged resources (databases & singularity images):

rm -rf resources
wget <archive_url>
tar -xvzf resources.tar.gz

---

Configuration

To enable rMLST support, configure API access to the PubMLST REST service by creating a file at config_files/.env with your credentials:

CONSUMER_KEY=""
CONSUMER_SECRET=""
TEST_REST_URL="https://rest.pubmlst.org/db/pubmlst_rmlst_seqdef_kiosk/schemes/1"
SESSION_REST_URL="https://rest.pubmlst.org/db/pubmlst_test_seqdef"
TEST_WEB_URL="https://pubmlst.org/bigsdb?db=pubmlst_test_seqdef"
ACCESS_TOKEN_PATH="<path_for_access_token>"

• Request an API key pair from PubMLST and fill in CONSUMER_KEY and CONSUMER_SECRET.
• See official documentation for details.
• If rMLST is not configured, species will be inferred using Kraken2 with the MiniKraken database.
• If rMLST is configured, species will be determined by a voting system that gives precedence to rMLST over Kraken2.

The config_files/yaml/config_modular_local.yaml file can be customized—for example, to define a custom location for the scratch directory used for intensive I/O operations.

• Relative paths should be specified with respect to the root of the Ardetype repository folder.
• Absolute paths are also supported, as long as they are accessible from the machine running the pipeline.

---

Usage

Input Structure Examples:

• isolate stands for sample ID.
• Having _ as part of sample ID may cause unexpected behavior.
• isolate in total must shorter than 27 characters, otherwise chewbbaca (cgMLST) will fail.

illumina/
├── isolate_R1_001.fastq.gz
└── isolate_R2_001.fastq.gz

ont/
└── isolate_ONT.fastq.gz

pre-assembled/
└── isolate_contigs.fasta

pacbio/
└── isolate_contigs.fasta

Typical Usage Examples:

# Illumina/Illumina+Nanopore data for each isolate
python ardetype.py -i <batch_path> -m all

# Nanopore data
python ardetype.py -i <batch_path> -m all -ont

# Pre-assembled/PacBio (_contigs.fasta) data
python ardetype.py -i <batch_path> -m all -fa

# Merge batches (Illimina + Illumina; Nanopore + Nanopore; Hybrid + Hybrid; Fasta + Fasta) - add -fa or -ont flag for fasta/Nanopore data
python ardetype.py -m merge --merge_from <batch_path1> <batch_path2> <batch_path3> -o <target_batch_path>

# Rerun all rules
python ardetype.py -m all -i <batch_path> -f

# Rerun specific rules (example) - use rule names as defined in snakefiles
python ardetype.py -m all -i <batch_path> -fr resfinder plasmidfinder

Notes:

• It is recommended to use the screen utility to run the pipeline in the background, especially for long-running jobs or remote sessions.
• In merge mode, the folder specified by <batch_path> must contain a marker file indicating the sequencing technology used — either ILL_mark, ONT_mark, or HYB_mark.
  This file is automatically generated by the pipeline during initial processing.
• The -fr option does not validate whether the specified rules exist in the pipeline.
  Please double-check rule names (snakefiles/) and review the output to ensure that the expected steps are reprocessed correctly.
• Refer to config_files/json/specific_tool_map.json for details on each tool and the species it is applicable to.
• By default, live pipeline logs are saved in the Ardetype/.snakemake/logs/ directory.
• The pipeline will iteratively exclude samples that fail any Snakemake rule in any module more than --retry_times (default: 3).
  • All input and output files associated with excluded samples will be moved to the <batch_path>_failed directory.
  • If all samples are excluded, the pipeline will terminate.
  • Detailed error information can be found in the Ardetype/.snakemake/log/ and <batch_path>/logs/ directories.

Command-line options:

  -i, --input
      Path to directory that contains files to be analysed.
      Files are NOT parsed from subfolders.
      Check `config_files/json/module_data/patterns/inputs` for expected input file extensions for the `bact_core` module.

  -c, --config
      Path to a custom pipeline configuration file in YAML format.
      If not provided, a default config file (`config_files/yaml/config_modular_local.yaml`) will be used.
      The configuration defines global parameters and tool-specific settings used during execution.

  -o, --output_dir
      Directory where pipeline outputs will be saved.
      By default, it is set to the same value as `--input`, which is recommended for idempotent re-runs.
      In `merge` mode, this must be explicitly specified as the destination for combined results.

  -r, --retry_times
      Number of times Snakemake will attempt to rerun failed jobs.
      Helps recover from intermittent network or I/O errors. Default is 3.

  -n, --num_jobs
      Maximum number of Snakemake jobs to run in parallel (when executing on HPC).
      Default is 12.

  -cpu, --max_cpus  Maximum number of CPUs Snakemake is allowed to use to run the jobs on     
      local machine. Default is 6.

  -hpc, --hpc
      Enable HPC execution mode.
      With appropriate configuration, this allows submitting each pipeline module
      (via `subscripts/ardetype_jobscript.sh`, `--hpc`, and `-sm` flags)
      or per-sample jobs (via `config_files/yaml/cluster.yaml` and `--hpc`) to a cluster scheduler (PBS by default).
      By default, the pipeline runs locally on the machine where the script is invoked.

  -sm, --submit_modules
      Run each pipeline module as a separate HPC job.
      Snakemake sub-jobs are not submitted to the cluster in this mode.
      Requires manual adjustment of `subscripts/ardetype_jobscript.sh` to control behavior.
      Effective only with `--hpc`.

  -f, --force_all
      Force rerun of all Snakemake rules across all pipeline modules,
      regardless of previous results or timestamps.

  -sp, --skip_packing
      Disable output packing.
      By default, output files are organized into per-sample folders.
      Use this flag to keep a flat output directory structure.

  -ont, --nanopore_only
      Run the pipeline on a folder containing exclusively Nanopore reads (with `_ONT.fastq.gz` extension).
      By default, the pipeline expects either Illumina data (`_001.fastq.gz`) or combined Illumina and Nanopore data.

  -mf MERGE_FROM [MERGE_FROM ...], --merge_from MERGE_FROM [MERGE_FROM ...]
      One or more paths to folders containing ARDETYPE output to be merged.
      Allowed combinations: Hybrid + Hybrid, Illumina + Illumina, Nanopore + Nanopore, Fasta + Fasta.

  -fr FORCE_RULES [FORCE_RULES ...], --force_rules FORCE_RULES [FORCE_RULES ...]
      Specify one or more Snakemake rule names (check `snakefiles/`) to force rerun within the pipeline.


required arguments:
  -m, --run_mode
      Selects pipeline execution mode:

        all    - Run the complete pipeline on raw FASTQ files, including all analysis modules:
                   `bact_core`, `bact_shell`, `bact_tip`, `bact_shape`.

        merge  - Combine outputs from previous ARDETYPE runs (see `--merge_from`),
                   then generate unified reports under `--output_dir`.

Report summary

File	Description
aquamis_qc_report.csv	comparing QC values against threshold defined in AQUAMIS pipeline
ardetype_report.csv	QC and truncated species-specific typing information
agrvate_report.csv	agr typing for S.aureus
amrfp_mutation_report.csv	Detecting AMR markers in bacterial NGS data.
chewbbaca_qc_report.csv	cgmlst profile quality using chewbbaca & Ridom schema
ectyper_report.csv	serotyping for E.coli
emmtyper_report.csv	emm-type and emm-cluster prediction for S.pyogenes
harmonized_resistance_profile.tsv	combined resfinder and RGI/CARD predictions
hicap_report.csv	serotyping for H. Influenzae
kleborate_report.csv	serotyping, resistance and virulence profiles for K.pneumoniae
kraken2contigs_report.csv	all hits with kraken2/minikraken applied to contigs
kraken2reads_report.csv	all hits with kraken2/minikraken applied to contigs
legsta_report.csv	sequence-based typing (SBT) for L. Pneumophila
lissero_report.csv	serotyping for L.monocytogenes
lrefinder_report.csv	predicting linezolid resistance in Enterococci faecalis and E. faecium
meningotype_report.csv	serotyping for N.meningitidis
mobtyper_contig_summary.csv	Detection and reconstruction of mobile genetic elements from bacterial NGS data
mobtyper_summary.csv	Detection and reconstruction of mobile genetic elements from bacterial NGS data
plasmidfinder_summary.csv	Detecting mobile genetic elements in bacterial NGS data (contigs)
pointfinder_report.csv	Detecting AMR-associated mutations in bacterial NGS data (contigs)
quast_report.csv	Quality control of genomes assembled from short-read NGS data
resfinder_mobtyper_mapping.csv	AMR-MGE link prediction from MOBSuite and Resfinder based on contig overlap
resfinder_pheno_table_gathered.csv	phenotypic resistance predictions by resfinder
seqsero2_report.csv	serotyping for S.enterica
seroba_report.csv	serotyping for S.pneumoniae
shigatyper_report.csv	serotyping for Shigella
sistr_report.csv	serotyping for S.enterica
software_log.csv	software/database versions
spatyper_report.csv	spa typing for S.aureus
stecfinder_report.csv	Shigatoxin typing for E.coli
virulencefinder_summary.csv	Detecting virulence genetic elements in bacterial NGS data (including full VFDB database)
pf_abaumannii_report.csv	Acinetobacter baumannii MGA search results
plassembler_report.csv	Hybracter plasmid assembly results - requires ONT data
plsdb_report.csv	PLSDB similarity search results for MOB-Suite constructed plasmid sequences