Metaxa

Metaxa is a deep learning–based classifier for metagenomic data that predicts taxonomic labels at the species and genus levels.

Installation

1. Clone the repository:

   git clone https://github.com/lbcb-sci/metaxa.git
   cd metaxa

2. Install dependencies:

   pip install -r requirements.txt
   pip install flash-attn --no-build-isolation

Note: The provided requirements.txt includes dependencies for CUDA 12.4. If you're using a different CUDA version, please install the appropriate version of torch, torchvision, and torchaudio manually, and remove their entries from requirements.txt before running pip install.

3. Download model:

   wget -O metaxa-model.v0.1.ckpt  https://zenodo.org/records/15062544/files/metaxa-model.v0.1.ckpt?download=1

Usage

Once installed, you can run Metaxa from the command line:

python metaxa/inference.py -c model.ckpt -d cuda:0 -b 1024 --n_workers 16 -o output.tsv reads.fastq

Arguments

Argument	Description	Example
--checkpoint, -c	Path to model checkpoint	-c checkpoint.ckpt
--device, -d	Device to run inference on	-d cuda:0
--batch_size, -b	Batch size	-b 1024
--n_workers	Number of data loading workers	--n_workers 16
--output, -o	Path to output classification file	-o output.tsv
	Input FASTQ/A file with sequences to classify	reads.fastq

Output

The output is a TSV file where each row contains:

• Read identifier (read_id),
• Predicted species-level taxonomic ID (species_taxid),
• Predicted genus-level taxonomic ID (genus_taxid).

Acknowledgements

This research is supported by the Singapore Ministry of Health’s National Medical Research Council under its Open Fund – Individual Research Grants (NMRC/OFIRG/MOH-000649-00).

Citing Metaxa

Coming soon.