How to separate hypervariable regions sequenced with 16s Metagenomics Ion torrent

Hello to all.
First time writing on the forum and trying to analyze with qiime2 data generated with ion Torrent after using 16s Metagenomics kit.

As most of you will know, the 16s metagenomics kit amplifies 7 different regions of the 16S gene into 6 amplicons. So after downloading my data I have a FASTQ with a mix of all the single-end reads of all the diferent regions.

In order to compare my data with bibliography, I want to analyze only the sequences that correspond to the v3 region, Can anyone give me any clues to separate the sequences according to the region from which they come from? Maybe one of you has done it before.

I´m thinking on trying to separate them by aligning with the whole 16s gene as reference, or an anyone give me any clues to separate the sequences according to the region from which they come from? maybe one of you has done it before.

I can think of trying to separate them by aligning them with the whole 16s gene as a reference or try to filter them with the used primers but I do not have Bioinformatic experience and I dont know programs, plugins or resources to do this job.

Thanks to all

Hi Mikel,

I don’t know if you have already figured out a solution, but take a look at this paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4734828/ and scroll down to the section “Separating reads into different variable regions” . It may help you ? What do you think @ebolyen ?

Best,
Lauren

1 Like

@Mikgar,
Sorry it looks like nobody ever answered your question — I think it came in during a busy travel period for the moderators so must have slipped past our radar.

If you want to isolate V3 regions prior to denoising, you can use qiime cutadapt trim-paired or trim-single with the --p-discard-untrimmed option… though that will only work if the primer is in the read.

If you want to isolate after denoising (or reads are not in the primer so you need to), you can use qiime quality-control exclude-seqs to align against sequences of interest (like V3 regions extracted from your reference database with qiime feature-classifier extract-reads)

You could also do this outside of QIIME 2 (e.g., before importing)… that would be the way to do it if (a) you want to separate before denoising/OTU clustering and (b) primers are not in the reads. VSEARCH is the tool I’d recommend for aligning fastq sequences against a reference database, and @Lauren’s suggestion would work as well.

@Lauren thanks for responding, sharing a solution, and bringing this question back to our attention!