This repository contains a Nextflow pipeline for detecting Biosynthetic Gene Clusters (BGCs) in metagenomic assemblies using the tools antiSMASH, GECCO, and SanntiS. This pipeline is designed to be run on EBI_SLURM.

Table of Contents

• Installation
• Usage
• Configuration
• Input Files
• Output Files
• Output Directory Structure
• Pipeline Structure
• Contributing
• License
• Acknowledgments

Installation

Clone the Repository

git clone https://github.com/Finn-Lab/BGCdetection_pipeline.git
cd BGCdetection_pipeline
python -m pip install -e .

Usage

Test

nextflow run main.nf -profile debug --input test/files/dummy_input.csv --outdir TEST_OUTPUT

Full Run

To run the pipeline on a larger dataset, provide an appropriate configuration and input file:

nextflow run main.nf -profile ebi --input <path_to_input_csv> --outdir <output_directory>

Parameters

• --input: Path to the CSV file containing input information.
• --outdir: Directory where the output files will be stored.

Options

• --process: To run specific processes. Provide processes names separated by comma, e.g. BUILD_GBK,ANTISMASH. If the argument is not provided, all processes will be run.

Configuration

The pipeline uses various configuration files to customize execution for different environments:

• conf/base.config: Base configuration for most high-performance compute environments.
• conf/ebi_codon.config: Specific configurations for the EBI Codon cluster.
• conf/modules.config: Configuration for individual modules.
• conf/test.config: Configuration for running minimal tests.

Input Files

The input CSV file should contain the following columns:

• PREFIX: Prefix for the output files, typically formatted as "{ASSEMBLY_FILE_NAME}_{md5sum(RESULT_DIRECTORY)}".
• GBK_GZ: Input file for the analysis (Absolute path). Genebank file with protein and nucleotide sequneces. Best if output of assembly_extraction_pipeline.

Example:

PREFIX,GBK_GZ
ERZ6863740_FASTA_5d9374cdf7a9f3b3ee89d860a60abe88,/home/User/ERZ6863740_FASTA/ERZ6863740_FASTA.gbk.gz
ERZ9863740_FASTA_9d9374cdf7a9f3b3ee89d860a60abe88,/home/UserERZ6863740_FASTA/ERZ9863740_FASTA.gbk.gz

Output Files

The pipeline generates various output files in the specified output directory:

• GFF files with BGC annotations from antiSMASH, GECCO, and SanntiS.
• TSV files with InterProScan results.

Output Directory Structure

The output directory structure is defined using Nextflow's publishDir directive. The output GenBank files will be organized in subdirectories based on the prefix:

<outdir>/
  ├── <last_two_characters_of_prefix>/
  │   └── <full_prefix>/
  │       └── <analysis>/
  │           └── *.gff.gz|*.IPS.tsv.gz

Pipeline Structure

The pipeline is composed of several Nextflow processes and modules:

• EXTRACT_FAA_GZ: Generates protein FASTA file to be used in interproscan.
• ANTISMASH: Runs antiSMASH for BGC detection.
• INTERPROSCAN: Runs InterProScan for protein domain prediction.
• SANNTIS: Annotates BGCs using neural networks trained on InterPro signatures.
• GECCO_RUN: Runs GECCO for BGC prediction.

Contributing

We welcome contributions to improve this pipeline. Please fork the repository and submit a pull request.

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Acknowledgments

This pipeline is a modification of mettannotator. Special thanks to the MGnify team for their work on the original pipeline.