Suggestions for assembling paired end reads without making contigs

[The Holo Scene]

Hi!

I'm looking for a program that can assemble paired end (100bpx2 HiSeq) reads without making contigs. The reads are from amplicons that vary from 49bp to 154bp; the amplicons are from plants and animals. I've already removed the adapters with Cutadapt. I've tried Pandaseq, but it misses a lot of reads. Coperead expects the reads to be the same length. Trimmomatic wasn't stringent enough for removing adapters, so its out too .

If anyone could suggest a program or even a perl script that will put the read 1 and read 2 together, preferably while using the phred scores and give me a fastq file, that would be lovely.

[SNPsaurus]

I like SeqPrep (https://github.com/jstjohn/SeqPrep). Not sure if it is what you want. I would refer to the merging of two reads into a longer sequence as making a contig. If you want to take overlapping paired reads and convert them into a joined sequence, then SeqPrep will do it.

[The Holo Scene]

This looks like what I am looking for, I will try it, thanks! By contigs, I just meant longer contigs of multiple reads. I do want the two paired reads to become one tiny contig.

[JackieBadger]

FLASH is very good at this, easy to use, and takes into account quality scores.

http://ccb.jhu.edu/software/FLASH/

[The Holo Scene]

I''l try Flash. SeqPrep is giving me merged sequences, but is losing the quality scores in the middle (all ]]]]]]] in the middle)

[SNPsaurus]

The ]]] quality scores are actually extremely high scores, reflecting the small probability of an error when both reads agree.

[The Holo Scene]

Oh, I'm an idiot.

[SNPsaurus]

Since we don't normally get Phred scores so high on next-gen sequencers, my brain also auto-calls any punctuation mark quality scores as junk. I think I had the same reaction as you when I first saw the output!