Dereplication with fastq out file

[Stephanie Schnorr]

Dear vsearch community,

Just a simple question, but is it possible to input and output a fastq file for the dereplication command?

I find that I can input fastq, but it defaults to fasta out, and I cannot override this by using the fastqout option flag. I really need to keep my quality scores after derep.

Any suggestions or comments in case I have overlooked something?

Thanks a bunch!

stephanie

[Torbjørn Rognes]

Hi

Sorry, but there is currently no command to dereplicate FASTQ files, only FASTA files. It is not obvious how to handle the quality scores that may differ between two entries with the same sequence.

This issue has been raised earlier:

https://github.com/torognes/vsearch/issues/348

I'll consider adding the fastx_uniques command to vsearch in the future.

- Torbjørn

[Stephanie Schnorr]

Hi ok thank you for clarifying! I see the issue, yes. Maybe instead of the usearch strategy of averaging q-scores it makes sense to take the highest q-score values for each position since if these sequences are truly clones, then the bases should be resolved at the highest score. I don't know, pontificating way outside of my wheelhouse.

Ah ha I did not look at the github site for the issue. Thanks very much for the fast response! If you do consider adding fasx_uniques to vsearch then that could be very useful I think to many.

Best,

Stephanie

[Torbjørn Rognes]

Hi

I am happy to inform you that the fastx_uniques option is now available in vsearch version 2.20.0 which has just been released:

https://github.com/torognes/vsearch/releases/tag/v2.20.0

By default, it will use the quality score corresponding to the average error probability for each base, for compatibility with usearch. However, I've added the option fastq_qout_max to use the best quality score in each position instead.

Enjoy!

[Stephanie Schnorr]

Fantastic, thanks so much for creating and implementing this feature!