Average Nucleotide Identity (ANI) analysis calculates the percentage of nucleotide identity among the supplied nucleotide sequences. It produces a square matrix of the calculated values. This matrix allows for pairwise comparisons among the nucleotide sequences and helps determine their similarities.
There are several methods to calculate the ANI:
-
ANIb (based on BLAST algorithm)
-
ANIm (based on MUMmer algorithm)
-
TETRA (based on tetranucleotide signature occurrences)
There are several tools available for ANI analysis (Figueras et al., 2014). For example:
-
JSpecies (http://www.imedea.uib.es/jspecies) (multiple genome analysis)
-
Gegenees (http://www.gegenees.org/documentation.html) (Due to changes in NCBI Blast, Gegeneese may not function with Blast versions 2.10 and later. Blast version 2.9 should work OK)
-
EzGenome (http://www.ezbiocloud.net/ezgenome/ani) [online]
-
ANI calculator (http://enve-omics.ce.gatech.edu/ani/index) [online] (only two genomes per analysis)
-
Python3 package (pyani) (https://github.com/widdowquinn/pyani, https://pyani.readthedocs.io/en/latest/run_anim.html) based on (Richter and Rossello´-Mo´ra, 2009). Pyani uses MUMmer algorithm.
-
If you are working on HPC Cluster, load the required version of python
module load cray-python/3.10.10 -
Create a conda environment with the compatible version of python, matplotlib and pyani
conda create -n pyani_env python=3.10 "matplotlib<=3.7" "pyani>=0.2.12" -c bioconda -
Activate the ani environment
conda activate pyani_env -
Alternatively, you can use my conda environment for pyani
-
Download it HERE
-
conda env create -f pyani_env.yml -
Check it has been installed. Copy the following command and hit enter
average_nucleotide_identity.py --help
The above command will show the flags/options of the pyani program
-
Install dos2unix for changing file format
conda install conda-forge::dos2unix -
Check it has been installed. Copy the following command and hit enter
dos2unix -
Make two metadata files and name them as ‘classes.txt’ (Fig. 1) and ‘labels.txt’ (Fig. 2)
Figure 1. Classes
Figure 2. Labels
Note, the first column is the nucleotide sequences names
Second column is the label of the nucleotide sequences
-
Make a directory and name it as ‘ANI’ for example
-
Within the ‘ANI’ directory, make another directory and name it as ‘genomes’ for example
-
Keep all the nucleotide sequences, ‘classes.txt’ and ‘labels.txt’ in the ‘genomes’ directory
-
Check the line terminator of the ‘classes.txt’ and ‘labels.txt’ files as follows
file *.txt
- If ‘classes.txt’ and ‘labels.txt’ have CRLF (Windows) format, then convert them into Unix format as follows:
dos2unix *.txt
- Run the following command from the ‘ANI’ directory
average_nucleotide_identity.py -i genomes -o output_ANI --labels genomes/labels.txt --classes genomes/classes.txt -g --gmethod seaborn --gformat pdf,png -v -l ba_ANI.log
-
Note that you do not make the output directory beforehand. Otherwise, the command will exit with an ‘overwriting’ error
-
Command reference: widdowquinn/pyani#56
The final output of the ANI analysis looks like this (Fig. 3):
Figure 3. Results
Figueras, M.J., Beaz-Hidalgo, R., Hossain, M.J., Liles, M.R., 2014. Taxonomic Affiliation of New Genomes Should Be Verified Using Average Nucleotide Identity and Multilocus Phylogenetic Analysis. Genome Announc 2, e00927-14. https://doi.org/10.1128/genomeA.00927-14
Richter, M., Rossello´-Mo´ra, R., 2009. Shifting the genomic gold standard for the prokaryotic species definition | Proceedings of the National Academy of Sciences. PNAS 106, 19126–19131. https://doi.org/10.1073/pnas.0906412106