Introduction metaBIOMx

The metagenomics microbiomics pipeline is a best-practice suite for the decontamination and annotation of sequencing data obtained via short-read shotgun sequencing. The pipeline contains NF-core modules and other local modules that are in the similar format. It can be runned via both docker and singularity containers.

Pipeline summary

The pipeline is able to perform different taxonomic annotation on either (single/paired) reads or contigs. The different subworkflows can be defined via --bypass_<method> flags, a full overview is shown by running --help. By default the pipeline will check if the right databases are present in the right formats, when the path is provided. If this is not the case, compatible databases will be automatically downloaded.

For both subworkflows the pipeline will perform read trimming via Trimmomatic and/or AdapterRemoval, followed by human removal via Kneaddata. Before and after each step the quality control will be assessed via fastqc and a multiqc report is created as output. Then taxonomy annotation is done as follows:

Read annotation - paired reads are interleaved using BBTools. - MetaPhlAn3 and HUMAnN3 are used for taxonomy and functional profiling. - taxonomy profiles are merged into a single BIOM file using biom-format.

Contig annotation - read assembly is performed via SPAdes. - Quality assesment of contigs is done via Busco. - taxonomy profiles are created using CAT. - Read abundance estimation is performed on the contigs using Bowtie2 and BCFtools. - Contigs are selected if a read can be aligned against a contig and a BIOM file is generated using biom-format.

Installation

[!NOTE] Make sure you have installed the latest nextflow version!

Clone the repository in a directory of your choice: bash git clone https://github.com/CMG-GUTS/metabiomx.git

The pipeline is containerised, meaning it can be runned via docker or singularity images. No further actions need to be performed when using the docker profile, except a docker registery needs to be set on your local system, see docker. In case singularity is used, images are automatically cached within the project directory.

Usage

Since the latest version, metaBIOMx works with both a samplesheet (CSV) format or a path to the input files. Preferably, samplesheets should be provided. bash nextflow run main.nf --input <samplesheet.csv> -work-dir work -profile singularity nextflow run main.nf --input <'*_{1,R1,2,R2}.{fq,fq.gz,fastq,fastq.gz}'> -work-dir work -profile singularity

📋 Sample Metadata File Specification

metaBIOMx expects your sample input data to follow a simple, but strict structure to ensure compatibility and allow upfront validation. The input should be provided as a CSV file where each entry = one sample with specified sequencing file paths. Additional properties not mentioned here will be ignored by the validation step.

Minimum requirement

• sample_id ➡ every entry must have a unique, non-empty sample identifier.
• No spaces are allowed in sample IDs — use underscores _ or dashes - instead.
• forward_read ➡ every entry must provide a path to an existing forward read FASTQ file (gzipped).
• If reverse_read is provided, forward_read must also be present. Example:

| sampleid | forwardread | reverseread | |-----------|---------------|--------------------| | sample1 | sample1R1.fastq.gz | sample1R2.fastq.gz | | sample2 | D0293271.fastq.gz | D0293272.fastq.gz | | S3 | L9283R1.fastq.gz | L9283R1.fastq.gz |

Properties and Validation Rules

🔹 Required properties

| Property | Type | Rules / Description | |--------------|--------|----------------------------------------------------------------------------------------------------| | sample_id | string | Unique sample ID with no spaces (^\S+$). Serves as an identifier. | | forward_read | string | File path to forward sequencing read. Must be non-empty string matching FASTQ gzipped pattern. File must exist. |

🔹 Optional property

| Property | Type | Rules / Description | |----------------|--------|----------------------------------------------------------------------------------------------------| | reverse_read | string | File path to reverse sequencing read. Same constraints as forward_read. Required if specified. |

Example cases

🔹 Read annotation

bash nextflow run main.nf \ --input <samplesheet.csv> \ # (optional) --bypass_trim \ # (optional) --bypass_decon \ --bypass_contig_annotation \ -work-dir work \ -profile singularity

🔹 Contig annotation

bash nextflow run main.nf \ --input <samplesheet.csv> \ # (optional) --bypass_trim \ # (optional) --bypass_decon \ --bypass_read_annotation \ -work-dir work \ -profile singularity

In case you only have assemblies and wish to perform contig annotation: bash nextflow run main.nf \ --input <samplesheet.csv> \ --bypass_assembly \ --bypass_read_annotation \ -work-dir work \ -profile singularity

Automatic database setup

The pipeline requires a set of databases which are used by the different tools within this workflow. The user is required to specify the location in where the databases will be downloaded. It is also possible to download the databases manually. The configure subworkflow will evaluate the database format and presence of the compatible files automatically. bash nextflow run main.nf \ --bowtie_db path/to/db/bowtie2 \ --metaphlan_db path/to/db/metaphlan \ --humann_db path/to/db/humann \ --cat_pack_db path/to/db/catpack \ --busco_db path/to/db/busco_downloads \ -work-dir <work/dir> \ -profile <singularity,docker>

Support

If you are having issues, please create an issue

Citations

You can cite the metabiomx using the following DOI: https://doi.org/10.48546/workflowhub.workflow.1787.5

An extensive list of references for the tools used by the pipeline can be found in the CITATIONS.md file.