I'm planning on using Oxford Nanopore MinION sequencer to test a script that I'm working on. However, I would not like to run multiple MinION runs to validate the tool. I'd like to simulate various minION runs and get raw reads and all pore events in a fast5 file. Do you know of any Oxford Nanopore (MinION) sequencing simulator that allows to get fast5 as output? I looked into NanoSim but it seems like there is no option to get fast5 as an output.
Using a simulator for example MinION output doesn't make sense when there are an abundance of freely-available FASTQ and FAST5 sequences for general purpose use (and especially for the purposes of testing out algorithms). Simulation of sequences can only ever express the components of DNA that are explicitly modelled, which is never going to fully represent reality.
As long as you're doing your own private research, you can use the FAST5 sequences from Clive Brown's own genome:
Bear in mind the license attached to that data:
License to use for research use only (i.e. non-commercial) governed by the law of England and Wales.
You agree to this License by downloading the data on this site whether in whole or in part.
You cannot re-distribute modified copies of these data, or derivative works, without my express written permission. Where such permission is granted, you must cite the original work and its Copyrights and you must include this license with any work that includes any of my genome.
For a more open license, you can use the nanopore-wgs reads here (derived from human cell line NA12878):
We encourage the reuse of this data in your own analysis and publications which is released under the Creative Commons CC-BY license. Therefore we would be grateful if you would cite the reference below if you do.
If you don't want human data, there's mouse parasite data here:
And E. coli data here:
... and if you just want 8h of sensor noise as a FAST5 file, I've got that here in one of my DropBox folders (as a ~70MB file):
Reminder: if you do ever present the results of analyses based on a data source, it's a good idea to acknowledge the source (even if not expected). If nothing else, it will let the creator know that someone cares about their data, and perhaps encourage them to release more in the future.
I wrote a tool that will rerun an experiment. So all you would need to do is just download an ONT dataset from ENA or something and you can just rerun it as many times as you want.