Do you think that DESEQ tool fit with microbiome metagenome analysis better than QIIME2 as some people recommended me to use this tool ? !
Some researchers recommended also the use of phyloseq and LEFSE for microbiome analysis. As i already used QIIME2 could you list some disadvantages of the use of the undermentioned tools compared with qiime
phyloseq
deseq
LEFSE
What are the arguments to provide about the usefulness of qiime2 in microbiome analysis rather than other tools ? Does qiime 2 provide appropriate corrections for multiple testing ?
Thanks
Hi @M_F,
The biggest difference is that QIIME 2 is intended as an end-to-end analysis platform for microbiome data, while DESEQ(2) is focused on the specific task of differential abundance testing. DESEQ won't, for example, help you with quality control of your data (e.g., as you'd do with qiime dada2 denoise-paired) or taxonomic annotation (e.g., qiime feature-classifier ...). phyloseq provides more functionality than DESEQ, but as far as I understand it hasn't been actively developed in several years so methods may be outdated.
QIIME 2 provides some functionality for differential abundance testing - we currently recommend ANCOM-BC. See the documentation of that here and here. We're also working on adding support for ANCOM-BC2 in this release cycle (due out in April 2025, but development versions may be available sooner).
It's possible to export data from QIIME 2, such that you could run differential abundance testing with DESEQ2 if you'd like, but ANCOM-BC has been shown to have considerably lower false positive rates than DESEQ2, so I don't recommend that. (Update: this has come up on the forum here.) This paper provides a really nice comparison of methods for differential abundance testing (thanks for the link @jwdebelius!).
QIIME 2 does integrate support for multiple testing correction where applicable. For example, in the context of differential abundance testing, this is included in the output of QIIME 2's ANCOM-BC wrapper. For example, see the "q-val" in this screenshot, which is derived from this figure in second bit of documentation that I linked you to above:
Another big selling point for QIIME 2 is that it integrates support for complete analysis reproducibility through its provenance tracking and Provenance Replay functionality. This can be hugely helpful to you when you're preparing your results for publication, or when you need to document your work for others. Oh, and QIIME 2 has an awesome forum, with many awesome moderators who are here to help as you work through your awesome analysis!
We can clearly recommend that users avoid using edgeR (a tool primarily intended for RNA-seq data) as well as LEfSe (without p-value correction) for conducting DA testing with 16S rRNA gene data.
