Merge multiple q2 artifacts prior to denoising

Hello Forum!

I have a relatively simple question but I haven’t been able to find a solution quite yet. Here is what I am trying to do: I have a set of samples all sequenced together, but four different spacers/adapters were used, creating four subsets of samples. The samples are demulitplexed, but still contain the spacer sequence. I imported the four subsets of samples into four separate artifacts with the type SampleData[PariedEndSequencesWithQuality]. To remove the unique spacers, I will have to run cutadapt on the four subsets, with the four output artifacts still being of the type ‘SampleData[PariedEndSequencesWithQuality]’ .

My question: is there a way to merge these four SampleData[PariedEndSequencesWithQuality] artifacts into one SampleData[PariedEndSequencesWithQuality] artifact? Therefore I can continue my pipeline straight into DADA2 denoising with the single artifact. Or, will I have to run denoising on the four artifacts separately and merge their resulting feature tables?

I guess my other option is to extract the .fastq.gz files from each artifact and re-import them into qiime2 as a single artifact. I would like to stay away from this if possible but could be the best option.

Details on my setup: using python3.6 on a server, with qiime2-2018.11 (only option on the server currently without setting up a docker container).

Thanks for your help!

Hi @Connor_Geraghty,

First, welcome to the forum!

It think integration of multiple regions is a holy grail in marker gene right now. The only thing I know that - in theory - does this is SMURF. It is umm… there’s an implementation ? in Matlab that’s basically a series of inhouse scripts strung together and called a day. Pretty dense and hard to use and not QIIME implemented. Ive run their example and it took me maybe… half and hour to profile the sample (with a substantial sacrifice to my swear jar when the FPM reached heights only previously matched when I tried to read SAS.)

But, you may also find this discussion on a similar issue interesting?


Hi Justine,

Thank you for your response! I should clarify these samples all had the same primer sets (V4, 515F/806R), but the way our sequencing core operates, they use four unique adapter/spacer pairs. I have to use cutadapt to trim each set separately, thus creating four artifacts that each contain 1/4 of the sample data. I am looking for a way to merge these four SampleData artifacts into a single artifact for downstream use.


Ahhh. Sorry for the confusion!

In this case, I would run Cutadapt on all of them, and then I think I’d denoise them separately, to generate a feature table and rep set. You can just merge those with the feature-table plugin. There’s a lot of answers here about merging if you search the forum and check out the FMT Tutorial


This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.