Bacteriophage MS2 Demo
You can try out Vanjari on the Bacteriophage MS2 (Emesvirus zinderi) genome.
wget "https://raw.githubusercontent.com/bloodhound-devs/vanjari/main/data/MS2.fasta" -O MS2.fasta
Then run the following command:
vanjari --input MS2.fasta --output-csv MS2-predictions.csv --image-dir MS2-images
This will classify the MS2 genome and save the predictions to a CSV file like this.
SequenceID |
Realm (-viria) |
Realm_score |
Subrealm (-vira) |
Subrealm_score |
Kingdom (-virae) |
Kingdom_score |
Subkingdom (-virites) |
Subkingdom_score |
Phylum (-viricota) |
Phylum_score |
Subphylum (-viricotina) |
Subphylum_score |
Class (-viricetes) |
Class_score |
Subclass (-viricetidae) |
Subclass_score |
Order (-virales) |
Order_score |
Suborder (-virineae) |
Suborder_score |
Family (-viridae) |
Family_score |
Subfamily (-virinae) |
Subfamily_score |
Genus (-virus) |
Genus_score |
Subgenus (-virus) |
Subgenus_score |
Species (binomial) |
Species_score |
NC_001417.2 |
Riboviria |
1.0 |
NA |
NA |
Orthornavirae |
1.0 |
NA |
NA |
Lenarviricota |
1.0 |
NA |
NA |
Leviviricetes |
1.0 |
NA |
NA |
Norzivirales |
1.0 |
NA |
NA |
Fiersviridae |
0.9998447 |
NA |
NA |
Emesvirus |
0.99968076 |
NA |
NA |
Emesvirus zinderi |
0.9996768 |
To create in image of the classification hierarchy, you can use the --image-dir
option:
vanjari --input MS2.fasta --output-csv vanjari-MS2.csv --image-dir vanjari-MS2-images
That will produce the following image in the vanjari-MS2-images
directory:

In this example, the model predicts the correct species with a confidence of more than 99%. In other cases where the confidence is lower, other possibilities will be shown in the image.
If you have Graphviz installed, the images will be rendered as PNG files, otherwise they will be saved as DOT files.
If you wish to render the images files as PDFs, SVGs, or other formats, you can use a command line option such as: --image-extension pdf
.
You can also run the VanjariFast model:
vanjari-fast --input MS2.fasta --output-csv vanjari-fast-MS2.csv --image-dir vanjari-fast-MS2-images
The same results can be produced programmatically:
from vanjari import Vanjari
vanjari = Vanjari()
vanjari(input=filename, output_csv="vanjari-MS2.csv", image_dir="vanjari-MS2-images")
Or you can use the VanjariFast model programmatically:
from vanjari.apps import VanjariFast
vanjari_fast = VanjariFast()
vanjari_fast(input=filename, output_csv="vanjari-fast-MS2.csv", image_dir="vanjari-fast-MS2-images")
Follow this link to launch a demo on Google Colab.