Metagenomics · Software
MetaCAT
Bins metagenome assemblies into genomes using coverage and composition, and links them with host traits.
What MetaCAT is
MetaCAT (Metagenome Clustering and Association Tool) is a bioinformatic package designed to reconstruct microbial genomes from metagenomic sequencing assemblies and assess their relationships with host phenotypes. It bins assembled contigs by evaluating sequence composition alongside cross-sample read coverage profiles.
Developed by C.C. Liu, S.S. Dong, J. Guo, and colleagues, MetaCAT provides a modular pipeline spanning contig clustering, bin quality checking, taxonomic placement, representative genome selection, abundance quantification, and sequence variant profiling. It is designed to facilitate the recovery of high-quality metagenome-assembled genomes across cohort-scale multi-sample datasets.
How it works
MetaCAT processes metagenomic contigs and alignment files through a series of dedicated commands. The initial stage identifies marker genes and seeds using FragGeneScan and HMMER search utilities, while sample BAM files are indexed and processed to quantify contig coverage across samples. Contigs are then partitioned into genome clusters based on coverage and sequence composition metrics.
Once candidate genome bins are formed, MetaCAT wraps external tools to curate and annotate them. It interfaces with CheckM2 to evaluate bin completeness and contamination, and uses GTDB-Tk for taxonomic classification. To dereplicate bins and establish reference sets, the representative module computes pairwise genome distances using Mash, FastANI, or skani. Cluster abundances and variants are subsequently calculated across samples. Matrix operations can optionally be accelerated on GPUs using CuPy.
Key facts
- Input formats
- Metagenomic FASTA assemblies and aligned reads in BAM format
- Key dependencies
- CheckM2, GTDB-Tk, FragGeneScan, HMMER, Mash, FastANI, skani, CuPy
- Quality assessment engine
- CheckM2
- Taxonomic classification engine
- GTDB-Tk
- Distance and dereplication engines
- Mash (default), FastANI, or skani
- GPU acceleration
- Supported via CuPy
- Runtime environment
- Python (tested on Python 3.12); pre-compiled binaries provided for x86_64 Linux and arm macOS
What you can use it for
- Metagenomic contig binning
- Partitioning de novo metagenomic assemblies into discrete microbial genome bins using multi-sample coverage depths and sequence characteristics.
- Genome quality estimation
- Evaluating completeness and contamination rates for reconstructed microbial bins via integrated CheckM2 runs.
- Taxonomic assignment
- Classifying reconstructed metagenome-assembled genomes into standard taxonomic lineages using GTDB-Tk.
- Genome dereplication and representative selection
- Identifying non-redundant representative genomes across multiple assemblies using Mash, FastANI, or skani.
- Abundance and variant profiling
- Quantifying the relative abundance and identifying genetic variants of reconstructed genomes across multi-sample cohorts to correlate with host traits.
Run it here
The research agent sets MetaCAT up on its machine the first time a task needs it, then runs it and reads the output back to you. Nothing to install on your side.
Try asking
“Bin my metagenome assembly into MAGs with MetaCAT using these coverage files.”
Strengths and limits
Good at
- Provides an integrated workflow from raw contig binning through to taxonomic classification, dereplication, and sample-level abundance profiling.
- Includes built-in binaries for FragGeneScan, HMMER utilities, and Mash on x86_64 Linux and arm macOS platforms.
- Allows users to choose between Mash, FastANI, and skani when selecting representative genomes.
- Supports GPU-accelerated computing through CuPy to improve processing speed.
Watch out for
- Full functionality depends on large external reference databases that must be downloaded separately for CheckM2 and GTDB-Tk.
- Platforms other than x86_64 Linux and arm macOS require manual compilation of underlying binaries including FragGeneScan, HMMER tools, and Mash.
- GPU acceleration requires appropriate NVIDIA CUDA runtime support and manual CuPy installation.
Questions
Can I run MetaCAT online without installing it?
The research agent sets MetaCAT up on its machine the first time a task needs it, then runs it and reads the output back to you. Nothing to install on your side.
What licence is MetaCAT under?
MetaCAT is published under its authors' own licence; the project page at https://github.com/liu-congcong/MetaCAT states it.
Can MetaCAT run without an NVIDIA GPU?
Yes. MetaCAT executes on standard CPUs by default, with CuPy and GPU usage serving as an optional acceleration method.
Which tools can MetaCAT use to choose representative genomes?
It uses Mash by default, but users can specify either FastANI or skani using the representative command's engine option.
Does MetaCAT bundle the databases needed for taxonomic classification and quality checks?
No. The GTDB reference data and CheckM2 Diamond databases must be acquired and configured independently.
Can MetaCAT be used solely for clustering contigs?
Yes. If downstream taxonomy or quality checks are not required, MetaCAT can be run independently without installing GTDB-Tk or CheckM2.
How to cite it
Liu, CC., Dong, SS., Guo, J. et al. MetaCAT enables reconstruction of high-quality microbial genomes and their association with host traits from metagenomic data. Nat Microbiol (2026). https://doi.org/10.1038/s41564-026-02472-7
Written from MetaCAT's own README and project page. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.