Skip to main content

pyani provides a package and script for calculation of genome-scale average nucleotide identity.

Project description


Whole-genome classification using Average Nucleotide Identity

pyani PyPi version pyani licence pyani TravisCI build status pyani coverage pyani Docker Pulls

pyani is maintained by:

and we are grateful to all who have contributed to this software:

Table of Contents


pyani is a program that calculates average nucleotide identity (ANI) and related measures for whole genome comparisons, and renders graphical summary output. Where available, it takes advantage of multicore systems, and can integrate with SGE/OGE-type job schedulers for the sequence comparisons.

This applies only to v0.2 of pyani.

The following scripts should be visible in your $PATH after installation:

  • that enables command-line ANI analysis.
  • that downloads publicly-available genomes from NCBI.
  • is a helper script required to run delta-filter on SGE/OGE systems.


The easiest way to install pyani v0.2 is to use conda or pip. conda is recommended for the simplest installation of third-party tool dependencies (mummer and BLAST/BLAST+).


You will need to install the bioconda channel, following instructions at Then, to create a new environment for pyani and install the program, issue the following command:

conda create --name pyani_env python=3.8 -y
conda activate pyani_env
conda install pyani


pip will install pyani and its Python dependencies, but not the third-party tools.

pip3 install pyani

Docker images

pyani v0.2 scripts are also provided as Docker images that can be run locally as containers on any operating system which supports Docker. To use these images, first install Docker. Then, to run the corresponding scripts, issue either:

docker run -v ${PWD}:/host_dir leightonpritchard/average_nucleotide_identity


docker run -v ${PWD}:/host_dir leightonpritchard/genbank_get_genomes_by_taxon


The -v ${PWD}:/host_dir argument enables the Docker container to see the current directory. Without this argument, the container will not be able to see your input files, or write output data.

Installing from the repository/source code

If you wish to install pyani v0.2 from source code, you will need to download this code from GitHub either directly, or by cloning the repository.

Both methods will create a directory called pyani which contains the source code for v0.2.


Please note that v0.2 is a MAINTENANCE-ONLY RELEASE and is found on the branch called version_0_2. Bugs and security issues will be fixed as we become aware of them, but no new features will be added. All new feature development now takes place under v0.3 (the master branch on the repository).

Obtain source code

You can obtain the source for pyani eithe by downloading the latest release as a compressed archive, or cloning the version_0_2 branch.

Direct download

Download the source as .zip or .tar.gz from:

Extract the archived file using the appropriate commands (or by double-clicking in your file explorer/finder):



tar -zxvf v0.2.10.tar.gz
Clone the repository using git

Download the current pyani repository with git clone, and change branch to version_0_2:

git clone
git checkout version_0_2

Installation From Source

Change directory to pyani, and use the python setuptools command, to install the package and scripts. This will not install the third-party tools BLAST/BLAST+ and mummer.

cd pyani
python build
python install

Third-party tools

Three alignment packages are required if you wish to use all of pyani's methods: mummer, BLAST+, and legacy BLAST.

NOTE: the legacy BLAST executables provided by NCBI will not run on macOS Big Sur; we do not provide executables for this tool.*

The simplest route to obtaining these tools is to use conda/bioconda:

conda install mummer blast legacy-blast -y

But they can also be installed by following instructions from the tools' own websites.

How To Run pyani


The script - installed as part of this package - enables straightforward ANI analysis at the command-line, and uses the pyani module behind the scenes.

You can get a summary of available command-line options with -h

$ -h
usage: [-h] [-o OUTDIRNAME] [-i INDIRNAME] [-v]
                                      [-f] [-s FRAGSIZE] [-l LOGFILE]
                                      [--skip_nucmer] [--skip_blastn]
                                      [--noclobber] [--nocompress] [-g]
                                      [--gformat GFORMAT] [--gmethod GMETHOD]
                                      [--labels LABELS] [--classes CLASSES]
                                      [-m METHOD] [--scheduler SCHEDULER]
                                      [--workers WORKERS]
                                      [--SGEgroupsize SGEGROUPSIZE]
                                      [--maxmatch] [--nucmer_exe NUCMER_EXE]
                                      [--blastn_exe BLASTN_EXE]
                                      [--makeblastdb_exe MAKEBLASTDB_EXE]
                                      [--blastall_exe BLASTALL_EXE]
                                      [--formatdb_exe FORMATDB_EXE]
                                      [--write_excel] [--subsample SUBSAMPLE]
                                      [--seed SEED] [--jobprefix JOBPREFIX]


Example data and output can be found in the directory test_ani_data. The data are chromosomes of four isolates of Caulobacter. Basic analyses can be performed with the command lines: -i tests/test_ani_data/ -o tests/test_ANIm_output -m ANIm -g -i tests/test_ani_data/ -o tests/test_ANIb_output -m ANIb -g -i tests/test_ani_data/ -o tests/test_ANIblastall_output -m ANIblastall -g -i tests/test_ani_data/ -o tests/test_TETRA_output -m TETRA -g

The graphical output below, supporting assignment of NC_002696 and NC_011916 to the same species (C.crescentus), and the other two isolates to distinct species (NC_014100:C.segnis; NC_010338:C. sp K31), was generated with the command-line: -v -i tests/test_ani_data/ \
    -o tests/test_ANIm_output/ -g --gformat png,pdf,eps \
    --classes tests/test_ani_data/ \
    --labels tests/test_ani_data/

ANIm percentage identity for Caulobacter test data ANIm alignment coverage for Caulobacter test data ANIm alignment length for Caulobacter test data ANIm alignment similarity errors for Caulobacter test data


The script, installed by this package, enables download of genomes from NCBI, specified by taxon ID. The script will download all available assemblies for taxa at or below the specified node in the NCBI taxonomy tree.

Command-line options can be viewed using:

$ -h
usage: [-h] [-o OUTDIRNAME] [-t TAXON] [-v]
                                        [-f] [--noclobber] [-l LOGFILE]
                                        [--format FORMAT] [--email EMAIL]
                                        [--retries RETRIES]
                                        [--batchsize BATCHSIZE]

For example, the NCBI taxonomy ID for Caulobacter is 75, so all publicly-available Caulobacter sequences can be obtained using the command-line:

$ -o Caulobacter_downloads -v -t 75 -l Caulobacter_downloads.log --email
INFO: Mon Apr 18 17:22:54 2016
INFO: command-line: /Users/lpritc/Virtualenvs/pyani3/bin/ -o Caulobacter_downloads -v -t 75 -l Caulobacter_downloads.log --email
INFO: Namespace(batchsize=10000, email='', force=False, format='gbk,fasta', logfile='Caulobacter_downloads.log', noclobber=False, outdirname='Caulobacter_downloads', retries=20, taxon='75', verbose=True)
INFO: Set NCBI contact email to
INFO: Creating directory Caulobacter_downloads
INFO: Output directory: Caulobacter_downloads
INFO: Passed taxon IDs: 75
INFO: Entrez ESearch with query: txid75[Organism:exp]
INFO: Entrez ESearch returns 29 assembly IDs
INFO: Identified 29 unique assemblies
INFO: Taxon 75: 29 assemblies
INFO: Assembly 639581: 271 contigs
INFO: Assembly 233261: 17 contigs
INFO: Assembly 575291: 48 contigs
INFO: Mon Apr 18 17:25:46 2016
INFO: Done.

NOTE: You must provide a valid email to identify yourself to NCBI for troubleshooting.

The number of attempted retries for each download, and the size of a batch download can be modified. By default, the script will attempt 20 download retries, and obtain sequences in batches of 10000.

Testing pyani

The pyani repository includes tests that can be run with pytest (including coverage measurement using pytest-cov) with the following command, executed from the repository root directory:

pytest -v

Method and Output Description

Average Nucleotide Identity (ANI)

This module calculates Average Nucleotide Identity (ANI) according to one of a number of alternative methods described in, e.g.

  • Richter M, Rossello-Mora R (2009) Shifting the genomic gold standard for the prokaryotic species definition. Proc Natl Acad Sci USA 106: 19126-19131. doi:10.1073/pnas.0906412106. (ANI1020, ANIm, ANIb)
  • Goris J, Konstantinidis KT, Klappenbach JA, Coenye T, Vandamme P, et al. (2007) DNA-DNA hybridization values and their relationship to whole-genome sequence similarities. Int J Syst Evol Micr 57: 81-91. doi:10.1099/ijs.0.64483-0.

ANI is proposed to be the appropriate in silico substitute for DNA-DNA hybridisation (DDH), and so useful for delineating species boundaries. A typical percentage threshold for species boundary in the literature is 95% ANI (e.g. Richter et al. 2009).

All ANI methods follow the basic algorithm:

  • Align the genome of organism 1 against that of organism 2, and identify the matching regions
  • Calculate the percentage nucleotide identity of the matching regions, as an average for all matching regions

Methods differ on: (1) what alignment algorithm is used, and the choice of parameters (this affects the aligned region boundaries); (2) what the input is for alignment (typically either fragments of fixed size, or the most complete assembly available).

  • ANIm: uses MUMmer (NUCmer) to align the input sequences.
  • ANIb: uses BLASTN+ to align 1020nt fragments of the input sequences
  • ANIblastall: uses legacy BLASTN to align 1020nt fragments of the input sequences
  • TETRA: calculates tetranucleotide frequencies of each input sequence

The algorithms takes as input correctly-formatted FASTA multiple sequence files. All sequences for a single organism should be contained in only one sequence file. Although it is possible to provide new labels for each input genome, for rendering graphical output, the names of these files are used for identification so it is best to name them sensibly.

Output is written to a named directory. The output files differ depending on the chosen ANI method.

  • ANIm: MUMmer/NUCmer .delta files, describing each pairwise sequence alignment. Output as tab-separated plain text format tables describing: alignment coverage; total alignment lengths; similarity errors; and percentage identity (ANIm).
  • ANIb and ANIblastall: FASTA sequences describing 1020nt fragments of each input sequence; BLAST nucleotide databases - one for each set of fragments; and BLASTN output files (tab-separated tabular format plain text) - one for each pairwise comparison of input sequences. Output as tab-separated plain text tables describing: alignment coverage; total alignment lengths; similarity errors; and percentage identity (ANIb or ANIblastall).
  • TETRA: Tab-separated plain text files describing the Pearson correlations between Z-score distributions for each tetranucleotide in each input sequence (TETRA).

If graphical output is chosen, the output directory will also contain PDF, PNG and EPS files representing the various output measures as a heatmap with row and column dendrograms. Other output formats (e.g. SVG) can be specified with the --gformat argument.

Developer notes

The pyani package is presented at GitHub under two main branches:

  • master is the source code underpinning the most recent/current release of pyani. It will (almost) always be in sync with the latest release found at The only time this code should not be in sync with the release is when there are modifications to documentation, or immediately preceding a release.
  • development is the current bleeding-edge version of pyani. It should (almost) always be in a working and usable condition, but may not be complete and/or some features may be missing or still under development.

Code Style and Pre-Commit Hooks

The source code for pyani is expected to conform to flake8 linting, and black code styling. These are enforced as pre-commit hooks using the pre-commit package (included in requirements.txt).

The black and flake8 hooks are defined in .pre-commit-config.yaml. Custom settings for flake8 are held in .flake8.

To enable pre-commit checks in the codebase on your local machine, execute the following command in the root directory of this repository:

pre-commit install


Unless otherwise indicated, all code is subject to the following agreement:

(c) The James Hutton Institute 2014-2019 (c) The University of Strathclyde 2019-2021 Author: Leighton Pritchard


Address: Leighton Pritchard Strathclyde Institute for Pharmacy and Biomedical Sciences University of Strathclyde 161 Cathedral Street Glasgow Scotland G4 0RE UK

The MIT License

Copyright (c) 2014-2019 The James Hutton Institute Copyright (c) 2019-2021 The University of Strathclyde

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.


Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pyani-0.2.12.tar.gz (298.5 kB view hashes)

Uploaded source

Built Distribution

pyani-0.2.12-py3-none-any.whl (59.9 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page