Bonito is a Quartznet-based Deep Learning basecaller for Oxford Nanopore long-read sequencing.
pip install ont-bonito
The input data for bonito is in the form of one or more .fast5 files containing the raw data output from the sequencer.
We provide some example .fast5 files from ONT sequencing of SARS-CoV-2 using the ARCTIC protocol, collected by the CADDE project. (Citation: http://virological.org/t/first-report-of-covid-19-in-south-america/409).
This is located in the data directory.
To extract the data use the following command:
tar -xvzf data/SP1-3-fast5.tar.gz
The bonito basecaller command is used to perform basecalling. The model is named dna_r9.4.1 and is intended for DNA sequencing data sequenced using the R9.4.1 pore. Bonito currently does not include a model for direct RNA sequencing. The directory containing one or more .fast5 files is given as input.
bonito basecaller dna_r9.4.1 data/SP1-3-fast5 > basecalls.fasta
If you have a turing or volta GPU the --half flag can be used to increase performance.
The output is saved in .fasta format. If multiple .fast5 files are supplied as input, all basecalled read sequences are present in the same output file.
head basecalls.fasta
>00e1904e-0c09-4673-be53-4d6e4ce9c997
ATCAGAATAGTGCCATGGGTGGCACGTTGAGAAGAATGTTAGTTTCTGGATTGAATGACCACATCTGGAACGCGTACGCGCAAACAGTCTGAAAGAAGCA
ATGAAATGAGCCACATCAAGCCTACAAGACAAGCCATTGCGATAGCAATTCCACCAGTGATCCAATTTATTCTGCAAACAGCAACCAAGCACAAAACAAG