Fastq header modification

Question

[Guillaume]

Hi everyone,

I hope this question has not been posted before, even if I looked for an answer in the forum I didn't find anything that fits my problem.

I need to modify the header of my fastq files (generated by MinION) in order to fit with the new storage of the fast5 files.

For example, the name is :

@ba66450b /root/**barcode09/15/**file.fast5

And I need :

@ba66450b /root/file.fast5

And the number can be 15 but also 20,21,25, etc..

Thanks in advance if you can help me !

G

[Ram]

Please use the formatting bar (especially the code option) to present your post better. I've done it for you this time.

[Guillaume]

OK thank you, I'll do that next time.

[Ram]

Is sed an option? If the number is the only variable, you can write a regex around it.

[Sej Modha]

How about following sed command?

   sed -e 's/\*\*barcode09\/[0-9]*\///g'

[JC]

Or perl:

perl -pe 's/\*\*barcode\d+\/\d+\/\*\*//' < file_in > file_out

[Ram]

The reason I asked "Is sed OK" and did not give sed/perl code was so OP could work on writing their own code.

[GenoMax]

Give up on that dream @Ram :) Not going to happen.

[Ram]

I care more about people providing ready-to-use code than people asking for such code. The latter depend on the former, and the former encourage the latter. Askers will always exist, providers need to create scarcity so askers turn to themselves.

[WouterDeCoster]

Many of us are guilty of handing out free drinks erm solutions. I agree that it would be nice that askers can learn, but often the formulation of their question already indicates they have no idea on how to solve their issue.

[Ram]

Hence, the pointers. Use tool X, with options y and z, focusing on solving the challenge around concept c. Use sed to substitute a regex focusing on the number part of the identifier.

[WouterDeCoster]

I wonder is pepsi OK?

[Ram]

I prefer

[Guillaume]

perl -pe 's/**barcode\d+\/\d+\/**//' < file_in > file_out

Nice, this one is working. Thanks

[Ram]

The reason I asked "Is sed OK" and did not give sed code was so OP could work on writing their own code.

[Guillaume]

I totally understand you want askers to be able to write their own code. However for people like me that don't have strong bioinformatics skills it's hard to understand the code, except if you give us details enough to explain the line. You will tell me "follow bioinformatics courses", yes you're right, I should do that.

Thanks a lot for your kind help

[Ram]

I understand that people might be just starting out, but one has to start _somewhere_. You could google for the code and adapt that code - that way, you'll learn both Google search as well as code tweaking. By copy pasting code here, how would you know what each part of the code does?

One can either claim to be a beginner and stay there with the code others provide, or they can work their way towards not being a beginner any more.

[WouterDeCoster]

But not everyone has the same ambitions - to become a real biofinformagician.

[Sej Modha]

+1 for Bioinformagicians :)

[Ram]

Fair point. OP seems like they're willing to learn, so I'm doing the socratic method spiel.