Assembling mini barcodes into full-length gene

lvogel · 09-04-2017, 07:34 AM

Hi,
I have metabarcoding reads from an Illumina MiSeq run. I usually use a primer set which amplifies a certain fragment of COI. Now, however, we have used three different primer sets which produce overlapping fragments. My question is, is it possible to first merge the paired ends from each primer, and then assemble the fragments together, and then continue with the usual metabarcoding pipeline? I find much information online about merging paired-ends, but not about incorporating longer contigs.

I'm a grad student, instructed to investigate this, and I wanted to ask around here first, before I waste a lot of time trying to do this if my efforts are misguided. Thanks.

GenoMax · 09-04-2017, 10:59 AM

Provided your primers have covered the gene fully you could assemble the reads (after merging the pairs post demultiplexing/removal of barcodes).

Should be easy enough to test. Use bbmerge.sh (for merging the reads) and tadpole.sh/seal.sh to assemble the data. Using bbduk.sh (for trimming) would be optional in case you have adapter contamination in your data. All tools part of BBMap suite. They all have individual threads.

lvogel · 09-05-2017, 12:57 AM

Thanks for the reply. Sounds good. I'll give it a try, and post again if I have problems.

lvogel · 09-12-2017, 07:38 AM

Dear GenoMax (or anyone),
Is there a reason why assembly should be done at the step where you said (after PE merging and removal of primers/tags), as opposed to waiting until the end, when we have already clustered the reads into OTUs (98% similarity), and assembling those? There's been a discussion about this in my lab, and I feel that the way you suggested is correct, but I don't know of any concrete justification as to why.
Thanks for all your assistance.

GenoMax · 09-12-2017, 07:45 AM

I assume your minibarcodes are tagging species specific primers that allow you to bin the reads first. So once you bin the three products for each species you would trim the barcodes/merge the reads and then assemble the full gene. Is that interpretation correct?

lvogel · 09-12-2017, 07:54 AM

Our primers aren't species-specific. We use the same 3 primer sets for all of the mixed samples of insects. We don't know what the species are likely to be until BLASTing at the end.

GenoMax · 09-12-2017, 08:00 AM

Not what I was imagining then. If the tags/extraneous sequence present is not interfering with your OTU assignments then you could do the trimming/assembly later.

Are you at least merging the reads first to get the longest possible representation before going on with the analysis?

lvogel · 09-12-2017, 11:37 AM

OK, thanks. I merge the paired-ends as a first step, yes, if that's what you mean.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PubMed: Evaluation of PacBio sequencing for full-length bacterial 16S rRNA gene class	Newsbot!	Literature Watch	0	11-16-2016 03:00 AM
Does the full length sequencing using Iso-Seq get the real full length ?	huan	Pacific Biosciences	5	12-22-2015 11:28 PM
Extracting a full length gene from illumina reads?	lordjonwald	De novo discovery	0	10-27-2014 08:01 AM

09-04-2017, 07:34 AM	#1
lvogel Junior Member Location: Western Europe Join Date: Nov 2013 Posts: 9	Assembling mini barcodes into full-length gene Hi, I have metabarcoding reads from an Illumina MiSeq run. I usually use a primer set which amplifies a certain fragment of COI. Now, however, we have used three different primer sets which produce overlapping fragments. My question is, is it possible to first merge the paired ends from each primer, and then assemble the fragments together, and then continue with the usual metabarcoding pipeline? I find much information online about merging paired-ends, but not about incorporating longer contigs. I'm a grad student, instructed to investigate this, and I wanted to ask around here first, before I waste a lot of time trying to do this if my efforts are misguided. Thanks.

09-05-2017, 12:57 AM	#3
lvogel Junior Member Location: Western Europe Join Date: Nov 2013 Posts: 9	Thanks Thanks for the reply. Sounds good. I'll give it a try, and post again if I have problems.

09-04-2017, 10:59 AM	#2
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,087	Provided your primers have covered the gene fully you could assemble the reads (after merging the pairs post demultiplexing/removal of barcodes). Should be easy enough to test. Use bbmerge.sh (for merging the reads) and tadpole.sh/seal.sh to assemble the data. Using bbduk.sh (for trimming) would be optional in case you have adapter contamination in your data. All tools part of BBMap suite. They all have individual threads.

09-12-2017, 07:38 AM	#4
lvogel Junior Member Location: Western Europe Join Date: Nov 2013 Posts: 9	Dear GenoMax (or anyone), Is there a reason why assembly should be done at the step where you said (after PE merging and removal of primers/tags), as opposed to waiting until the end, when we have already clustered the reads into OTUs (98% similarity), and assembling those? There's been a discussion about this in my lab, and I feel that the way you suggested is correct, but I don't know of any concrete justification as to why. Thanks for all your assistance.

09-12-2017, 07:45 AM	#5
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,087	I assume your minibarcodes are tagging species specific primers that allow you to bin the reads first. So once you bin the three products for each species you would trim the barcodes/merge the reads and then assemble the full gene. Is that interpretation correct?

09-12-2017, 07:54 AM	#6
lvogel Junior Member Location: Western Europe Join Date: Nov 2013 Posts: 9	Our primers aren't species-specific. We use the same 3 primer sets for all of the mixed samples of insects. We don't know what the species are likely to be until BLASTing at the end.

09-12-2017, 08:00 AM	#7
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,087	Not what I was imagining then. If the tags/extraneous sequence present is not interfering with your OTU assignments then you could do the trimming/assembly later. Are you at least merging the reads first to get the longest possible representation before going on with the analysis?

09-12-2017, 11:37 AM	#8
lvogel Junior Member Location: Western Europe Join Date: Nov 2013 Posts: 9	OK, thanks. I merge the paired-ends as a first step, yes, if that's what you mean.