Skipping make.contigs

Hi,

Because my reads have additional barcodes and randomized nucleotides within read 1 and 2 before the V4 primer sites, I can’t use make.contigs within mothur to merge the reads.

  1. In theory I should be able to get around this by using the fastq.info command on my merged fastq reads to generate the fasta and qual files?
  2. Does this command also change the sequence identifier line, “:” to “_” or do I have to do this additionally?
  3. Finally I just create the group file myself by taking the changed sequence identifier names and adding sample name information?

Please correct my reasoning or point me to an existing command within mothur if I missed something.

Thanks,
Dino

Can you use N’s to precede your barcodes to represent random bases? I think you’re going to find that the make.contigs clean up will be much better than other alignment methods.

  1. In theory I should be able to get around this by using the fastq.info command on my merged fastq reads to generate the fasta and qual files?

Sure.

  1. Does this command also change the sequence identifier line, “:” to “_” or do I have to do this additionally?

You don’t really need to do that.

  1. Finally I just create the group file myself by taking the changed sequence identifier names and adding sample name information?

Should work.

Pat

Awesome, thanks.

I can just clip the initial few bases of both reads where the barcodes and random bases occur and then just use make.contigs but that kinda beats the whole purpose of using this method in the first place. The random bases of varying lengths are for improving sequencing performance, but are also used to uniquely tag template molecules before PCR so you can correct for amplification and sequencing errors later.
Here’s the paper describing the method Practical innovations for high-throughput amplicon sequencing - PubMed and pipeline MT-Toolbox: improved amplicon sequencing using molecule tags - PubMed accompanying it.

Yeah I’m familiar with the approach - we tried it and actually got much worse results than what we report in Kozich.