SEQanswers

Go Back   SEQanswers > Sequencing Technologies/Companies > Oxford Nanopore



Similar Threads
Thread Thread Starter Forum Replies Last Post
Nanopolish without reference genome jordi Bioinformatics 2 11-08-2017 05:27 AM

Reply
 
Thread Tools
Old 09-19-2017, 07:16 AM   #1
anotherSAM
Junior Member
 
Location: Paris

Join Date: Sep 2017
Posts: 5
Default Nanopolish 1D

I am working on 1D minion data, albacore 2.0 called, and attempting to use nanopolish on canu derived contigs.
I am trying to follow the github/readme files but find it hard.

https://github.com/jts/nanopolish

First I am under the impression nanopolish uses raw fast5 files in order to polish but the initial step requires indexing of fastq files (albaore 2.0 step)

Secondly for the mapping stage 'bwa mem...', the only options are for ont2d data.

When are the fast5 files used if they are?
and how can I use nanopolish on 1D data?

Thanks in advance

Last edited by anotherSAM; 09-20-2017 at 04:42 AM. Reason: typo
anotherSAM is offline   Reply With Quote
Old 09-25-2017, 03:28 PM   #2
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 799
Default

The 'ont2d' flag sets up BWA parameters to work for ONT data. It was named at a time when 2D data was similar accuracy to what 1D data is at the moment. You should be fine using Albacore 2.0-called reads with the stated workflow.

The indexing process links the Fast5 files to the FASTQ records. Nanopolish uses this index and the FAST5 files to correct the reads during the polishing step.
gringer is offline   Reply With Quote
Old 11-03-2017, 05:30 AM   #3
apredeus
Senior Member
 
Location: Bioinformatics Institute, SPb

Join Date: Jul 2012
Posts: 146
Default

I've used minimap2 from Heng Li to map the long reads before polishing - it's much faster and gives similar results. Bwa with ont2d works as well, just somewhat slower.

Nanopolish has changed their workflow starting from version 0.8 - you now have to index the fastq file to the fast5 location, and it's much easier after this and uses far less space.

After this, fast5 files are used to generate the consensus.
apredeus is offline   Reply With Quote
Old 11-03-2017, 11:49 AM   #4
gringer
David Eccles (gringer)
 
Location: Wellington, New Zealand

Join Date: May 2011
Posts: 799
Default

Nanopolish also now includes methylation in its correction model, which (according to Ryan Wick) has brought the nanopore-only consensus accuracy up to q30 for a bacterial genome.
gringer is offline   Reply With Quote
Old 11-03-2017, 07:45 PM   #5
cstack
Member
 
Location: Florida, US

Join Date: May 2017
Posts: 12
Default

Quote:
Originally Posted by anotherSAM View Post
I am working on 1D minion data, albacore 2.0 called, and attempting to use nanopolish on canu derived contigs.
I am trying to follow the github/readme files but find it hard.

https://github.com/jts/nanopolish

First I am under the impression nanopolish uses raw fast5 files in order to polish but the initial step requires indexing of fastq files (albaore 2.0 step)
Like an earlier user mentioned, the initial indexing step links each read in the fastq file with a fast5 record.

Prior to basecalling with albacore 2, you would use
Code:
nanopolish extract
to make the fastq-to-fast5 links; nanopolish would extract the base-called fastq sequences from fast5 files and would I think add comments to each fastq read indicating which fast5 file it came from. Having the fastq string stored in the fast5 file can cause the latter to grow significantly; and once you extract the fastq strings they become redundant in the fast5 files, and you end up wasting system storage. The newer version of nanopolish has the
Code:
index
subcommand that makes fastq-to-fast5 links so that using nanopolish might look like this:
  1. Run MinION without basecalling
  2. Call bases with albacore v2+ to fastq format
  3. Link fastq files to fast5 files with nanopolish index
  4. ... do additional stuff with nanopolish ...

Quote:
Originally Posted by anotherSAM View Post
Secondly for the mapping stage 'bwa mem...', the only options are for ont2d data.

When are the fast5 files used if they are?
and how can I use nanopolish on 1D data?

Thanks in advance
I think that the fast5 files are used when you run the
Code:
nanopolish variants --consensus
command. I think that you provide the fastq file as an argument, and nanopolish will check whether index files exist for those reads (usually same base filename with some new extension added), similar to the files generated by
Code:
bwa index
. The index files provide an explicit filesystem mapping between each read in the fastq file and a fast5 file. I've never tried, but I think that nanopolish will fail if you move the fast5 files to another location in the filesystem after you've indexed the reads. If you did that, I think you'd have to re-index.

hope that helps
cstack is offline   Reply With Quote
Reply

Tags
canu, nanopolish, ont

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 01:46 AM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO