--fastq_mergepairs

[Susan Newman]

Hi,

I have a very large dataset consisting of 90 forward and 90 reverse paired sequences.  I can successfully use --fastq_mergepairs to merge a single pair.  I've done this for a small dataset of 10 samples then used cat to concatenate the individual output files into a single file of merged contigs.  I have been unable to get a wildcard command to process all of my files to work.  I am following the USearch protocol using the command:

vsearch --fastq_mergepairs *_R1_*.fastq --reverse *_R2_*.fastq --fastqout mergedAll.fq

This command generates the error:  Unrecognized string on command line (M-42-CD-A_S68_L001_R1_001.fastq)  This is the second sample in my dataset.  And the error, "segmentation fault (core dumped)"

Can you please provide the proper syntax for handling multiple files?  How do I use --relabel to preserve the sample information in the merged file?

Many thanks for your assistance.  I am switching from USearch to VSearch as this dataset is too large to be handled by USearch-Windows version.

Susan

[Frédéric Mahé]

Hi Susan,


vsearch can only process one file (or pair of files) at a time. You should loop over your list of files. For example:

for R1 in *_R1_*.fastq ; do
    vsearch \
        --fastq_mergepairs ${R1} \
        --reverse ${R1/_R1/_R2_} \
        --fastqout ${R1/R1_*/merged.fastq}
done


If you process your files one by one, you can freely use the --relabel option to mark your reads.

[Susan Newman]

Yes, that did the trick.  I found one small typo, a missing underscore    --reverse ${R1/_R1_/_R2_}\ 

 Thanks for helping out.

Susan

[aebra...@gmail.com]

Hello Guys;

I have a plan to use usearch by linux command but I didn't find any command related to usearch in linux. Is anyone can help me how to analyse my data with usearch command in linux?

Thanks

Aziz

[Torbjørn Rognes]

Hi Aziz,

This is a support forum for vsearch, not for usearch. 

You can find and download the vsearch software here: https://github.com/torognes/vsearch

Please follow the instructions about downloading and installation in the README.

For more information about usearch, look here: http://www.drive5.com/usearch/

Good luck!

- Torbjørn

[bern...@gmail.com]

Hi Frédéric,

I used --relabel with -fastq_mergepairs, indeed expecting that the reads would be renamed (using vsearch v2.4.3_osx_x86_64)

However, relabelling of the read IDs does not occur, they remain unchanged. For example, with --fastq_filter and the same input files, the readIds are changed.

Should --relabel work indeed with --fastq_mergepairs ? 

Best,

Bernd

[Frédéric Mahé]

Hi Bernd,

no the --relabel option does not work with the command --fastq_mergepairs. To get a list of the accepted options for each command, try vsearch -h

Best,

[jstil...@gmail.com]

Is there a way to do this (multiple sets of paired reads merged into a single file) in the mac version?

[Frédéric Mahé]

do you mean something like this?

for f in *R1.fastq ; do vsearch --fastq_mergepairs ${f} --reverse ${f/R1/R2} --fastqout - ; done > all_merged.fastq