Generic AMR: A general purpose pipeline for AMR analyses

Caution

This pipeline is still experimental. Use at your own risk.

Generic AMR is a pipeline that extracts features from bacterial genomes useful for antimicrobial resistance (AMR) prediction.

Quickstart

1. Install Docker Engine or Desktop and Git on your system.
2. Open a terminal, and run the following commands:

: # 🌎 Clone this repository
git clone https://github.com/aapashkov/generic-amr

: # 📌 Change into base directory
cd generic-amr

: # 📄 Place your genome accessions in accessions.txt

: # ⚡ Run pipeline (you might need to run it as docker-compose on some systems)
docker compose run --rm pipeline

Configuration

The accessions.txt file lists, one per line, the genome and/or read accessions to process and includes some example accessions to get started. These accessions may be any of the following:

• NCBI Genome/RefSeq identifiers. They start with GCA_ or GCF_. For example, GCA_000005845.2.
• ENA sequence assembly analysis identifiers. They start with ERZ. For example, ERZ3086155.
• BV-BRC identifiers. They start with BVBRC_. For example, BVBRC_170673.13.
• NCBI SRA identifiers. They start with DRR, ERR or SRR. For example, SRR33085175. The sequencing technology is inferred by a request to NCBI SRA API, but can be overwritten by appending to the accession a tab (\t) and the technology, which can be one of: ILLUMINA, ION_TORRENT, LS454, OXFORD_NANOPORE, PACBIO_SMRT.
• Private genomes. When including these, place them in the data/genomes directory, decompressed and ending with .fna. Then, add their basenames (i.e., without the .fna extension) to accessions.txt.
• Private reads. When including these, place them in the data/reads/raw directory, one folder per sample, with the following structure:

  BASE/BASE.fastq.gz   → single/unpaired reads
  BASE/BASE_1.fastq.gz → forward reads
  BASE/BASE_2.fastq.gz → reverse reads
  

  Read identifiers must match @BASE.X[.Y], where X is the read ID, and Y must be 1, 2, or 3, respectively corresponding to forward, reverse and unpaired reads. Single reads (without paired-end data) must not contain .Y. After doing this, add the read BASEnames (i.e., folder names) to accessions.txt, followed by a tab (\t) and their sequencing technology, which can be one of: ILLUMINA, ION_TORRENT, LS454, OXFORD_NANOPORE, PACBIO_SMRT.

Warning

When using private reads, their names MUST NOT start with GCA_, GCF_, ERZ, or BVBRC_, as those will always be treated as genomes that have to be downloaded (ignoring the reads even if they exist).

Create a .env file in the root directory of this repository storing pipeline configuration details. Currently, the following environment variables are supported:

• UID (int) - your user ID, defaults to root if not specified.
• GID (int) - your group ID, defaults to root if not specified.
• JOBS (int) - number of CPUs to use for parallelization, defaults to all available CPUs.

If you wish to connect to the Docker container to perform manual operations instead of running the pipeline, use the following command:

: # 📦 Connect to Docker container without running the pipeline
docker compose run --rm pipeline bash

Output description

All pipeline outputs are stored in the data directory. You should expect a file structure like the following (where {accession} is a single entry in accessions.txt):

├── 📁 data/                # Pipeline outputs
│   ├── 📁 annotations/     # Genome annotations
│   │   └── 📁 {accession}/
│   │       ├── 📄 {accession}.platon.*     # Platon plasmid predictions
│   │       ├── 📄 {accession}.prokka.*     # Prokka annotations
│   │       ├── 📄 {accession}.rgi.*        # RGI annotations
│   │       └── 📄 {accession}.sourmash.*   # Sourmash signatures and taxonomies
│   ├── 📁 databases/       # Reference databases
│   │   ├── 📁 platon/      # Platon database
│   │   └── 📄 gtdb.sbt.zip # Sourmash database with genomes from GTDB RS226
│   ├── 📁 genomes/         # Decompressed genomes with .fna extension
│   │   └── 📄 {accession}.fna
│   └── 📁 logs/            # Pipeline execution logs, one per accession
├── 📁 env/                 # Environment definitions
│   ├── 📄 apt.txt          # Debian's apt package manager dependencies
│   └── 📄 pip.txt          # Python's pip requirements file
├── 📄 compose.yml          # Docker Compose execution parameters
├── 📄 Dockerfile           # Docker build definition
├── 📄 LICENCE              # MIT license
├── 📄 Makefile             # Pipeline steps
└── 📄 README.md            # Documentation