How demultiplexing works in Qiime?

Hi everyone!
I’m working on my bachelor’s thesis and it’s about how a microbiome data analysis is performed. I’m trying to explain every step of the pipeline showing the different methods and how every process works in details. So It’s have been a bit difficult to explain in details how demultiplexing works. In a broad way, I know that the process does, like described in Qiime2 documentation. Can someone suggest me a reference to have a better understanding or explain me what the algorithm does with some details?

Thanks a lot for your time and answer.

QIIME 2 currently has two plugins that support demultiplexing reads --- one that works with EMP-formatted reads, the other is a cutadapt-based plugin. As for the cutadapt plugin, you would need to read the cutadapt source and docs to get an idea of how that works --- we aren't the developers of cutadapt, we just use it. As for the EMP demuxing, that one doesn't really have an "algorithm" - it just looks at two (single-end) or three (paired-end) files - one file has the barcodes, the other(s) have the reads. If the first record in the barcodes file matches a barcode in our sample metadata file, then we copy the associated read(s) to the demuxed file for that particular sample.


Thanks a lot for your answer and suggestion Matthew the example image looks great! I’ll try to read the sources to have a better understanding. As soon as I do, I’ll leave the comment here for everyone give a little contribution.


It might also be good to look at barcoding strategies in general, since these impact your demultiplexing. I think the EMP uses Golay 12.



Good point @jwdebelius! Also worth noting, the EMP-demux in q2-demux does not perform any type of Golay error correction at present (it is on our todo list!). Thanks!


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