Tool for non-overlapping pair-ends reads?

[Marato]

Hi, I'm pretty new with metagenomics and my lab is using Illumina sequencing for generating an amplicon for purposes similar to 16S (OTU assignment) but using another marker which is longer (~580bp). The problem is that after getting the reads from MiSeq (2x300), the quality goes down after the 150 cycle and many of the reads had to be trimmed. The biggest problem is that most of the ends don't overlap, and one single end is not giving enough for blasting. Is there any tool that I can use to merge both ends? (I know there is going to be a gap, but I'd like at least to concatenate the sequences).
Thanks

[danova]

It looks like you have a problem with your dataset. Illumina miseq chemistry for (2x300pb) is known to give unpredictable results. We have switched back to 2x250 as we had similar problems to yours with 16S amplicons.. My best advice would be to repeat your experiment with 2x250bp. We normally use bbmerge.sh from the bbmap suite to merge pair-ended reads, however if you have really bad data and given the known problems with the 2x300 chemistry problems i would repeat the experiment.
Best,
david

[Marato]

Thanks for your answer. Yes, we have considered to repeat the experiment but we want to try anyway with our current data until other funds arrive. Does bbmap work fine with non-overlapping pair-ends?

[GenoMax]

BBMap works with PE reads but if your experiment was designed in such a way that the reads should have overlapped then something may be wrong beyond the quality issue. Have you tried BBMerge (part of BBMap suite) to do the merging or do you use something else (FLASH)?

[danova]

Try with bbmerge:
bbmerge.sh in1=file.R1.fq.gz in2=file.R2.fq.gz out=file.merged.fq.gz outu1=files.unmerged.R1.fq.gz outu2=file.unmerged.R2.fq.gz merge=t mininsert=300 qtrim2=t minq=20

Play with some parameters and see if you can merge....

[Marato]

I was using interlacer tool in Galaxy but didn't work. Thanks, I'll try bbmerge then