Amira is an AMR gene detection tool designed to work directly from bacterial long read sequences. Amira makes it easy to reliably identify the AMR genes in a bacterial sample, reduces the time taken to get meaningful results and allows more accurate detection of AMR genes than assembly.
Amira leverages the full length of long read sequences to differentiate multi-copy genes by their local genomic context. This is done by first identifying the genes on each sequencing read and using the gene calls to construct a de Bruijn graph (DBG) in gene space. Following error correction, the reads containing different copies of multi-copy AMR genes can be clustered together based on their path in the graph, then assembled to obtain the nucleotide sequence.
Amira requires Python and three additional non-Python tools for optimal functionality:
- Python >=3.9,<3.13.
- Poetry to manage the Python dependencies.
- Pandora to identify the genes on each sequencing read.
- minimap2 for sequence alignment.
- racon for allele polishing.
Follow these steps to install Amira and its dependencies.
Open a terminal and run the following command to clone the repository and navigate into it:
git clone https://github.com/Danderson123/Amira && cd AmiraAmira’s dependencies are managed with Poetry. Install Poetry by running:
pip install poetryOnce Poetry is installed, use it to set up Amira’s dependencies:
poetry installAmira requires Pandora, minimap2 and racon. Follow the links below for instructions on building binaries for each tool:
After installation, make a note of the paths to these binaries as they will be required when running Amira.
Amira can be installed from PyPI by running:
pip install amira-amrAmira can then be run with:
amira --helpAmira can be run on the output of Pandora directly, or from JSON files listing the genes and gene positions on each sequencing read. Below are instructions and an example command for running Amira with the JSON files.
To run Amira from the JSON files, you can use this command. You will need to replace <PATH TO RACON BINARY> with the absolute path to the racon binary you made earlier and replace <PATH TO MINIMAP2 BINARY> with the path to the minimap2 binary.
python3 amira/__main__.py --pandoraJSON <PATH TO GENE CALL JSON> --gene-positions <PATH TO GENE POSITION JSON> --pandoraConsensus <PATH TO PANDORA CONSENSUS FASTQ> --readfile <PATH TO READ FASTQ> --output <OUTPUT DIRECTORY> --gene-path <AMR GENE REFERENCE FASTA> --phenotypes <ALLELE-PHENOTYPE MAPPING JSON> --racon-path <PATH TO RACON BINARY> --minimap2-path <PATH TO MINIMAP2 BINARY> --debug --cores <CPUS> --sample-reads --filter-contaminants
Some example JSON data can be downloaded from here. Amira can then be run using this command:
python3 amira/__main__.py --pandoraJSON test_data/gene_calls_with_gene_filtering.json --gene-positions test_data/gene_positions_with_gene_filtering.json --pandoraConsensus test_data/pandora.consensus.fq.gz --readfile test_data/SRR23044220_1.fastq.gz --output amira_output --gene-path AMR_alleles_unified.fa --phenotypes AMR_calls.json --racon-path <PATH TO RACON BINARY> --minimap2-path <PATH TO MINIMAP2 BINARY> --debug --cores <CPUS> --sample-reads --filter-contaminants
Pandora uses species-specific reference pan-genomes (panRGs) to identify the genes on each sequencing read. For Escherichia coli, a pre-built panRG can be downloaded from here. After installing Pandora, you can call the genes on your sequencing reads using this command:
pandora map -t <THREADS> --min-gene-coverage-proportion 0.5 --max-covg 10000 -o pandora_map_output <PANRG PATH> <PATH TO READ FASTQ>Amira can then be run directly on the output of Pandora using this command:
python3 amira/__main__.py --pandoraSam pandora_map_output/*.sam --pandoraConsensus pandora_map_output/pandora.consensus.fq.gz --readfile <PATH TO READ FASTQ> --output amira_output --gene-path AMR_alleles_unified.fa --minimum-length-proportion 0.5 --maximum-length-proportion 1.5 --cores <CPUS> --phenotypes AMR_calls.json --filter-contaminants --sample-readsFor additional options and configurations, run:
python3 amira/__main__.py --helpTBD
If you’d like to contribute to Amira, please follow these steps:
- Fork the repository.
- Create a new branch for your feature or bugfix (git checkout -b feature-name).
- Commit your changes (git commit -m "Description of feature").
- Push to the branch (git push origin feature-name).
- Submit a pull request.
This project is licensed under the Apache License 2.0 License. See the LICENSE file for details.
For questions, feedback, or issues, please open an issue on GitHub or contact Daniel Anderson.