Integrating microbiome package stub within dada2 tutorial?

[benjjneb]

Hi microbiome team, Over in dada2 land we are interested in linking our package with other excellent Bioconductor microbiome packages. The primary way we have done this so far is to integrate "stubs" into our tutorial, which is by far our most accessed piece of documentation. To date, we have added brief examples and links to the DECIPHER package for their IDTaxa method, and the phyloseq package for more general microbiome analysis.

Would the microbiome team be amenable to dada2 adding a microbiome package stub to our tutorial? Do you have any suggestions on what sort of Rmarkdown code might be effective? (it should probably show "importing" of dada2 workflow data into microbiome and maybe the generation of a plot?)

I'm also interested in any other cross-linking ideas you might have. I think you all are doing great work!

[antagomir]

Sure, excellent idea, and tutorials do play a key role in dissemination of these methods.

The microbiome package currently focuses on downstream analyses of phyloseq objects, and provides some functionality for routine analyses and visualizations that we have not found elsewhere (or had not at the time of implementation). Suitable stubs would then most naturally demonstrate how to perform certain analysis/visualization tasks based on data that comes out of the dada2 workflow. Some possible topics include core microbiota. prevalence, and rare taxa; analysis and visualization of bimodal/bistable groups; and additional tools for manipulating phyloseq objects (transformations, filtering). Perhaps we could start with one, and add more later. We can use some data set from dada2 tutorial as the starting point, and send some Rmd code to be added to the tutorial, or make a PR.

We are currently revising our tutorial, too. Perhaps we could more explicitly aim to cross-link these resources.

[benjjneb]

OK that should be easy to add then. We already import the dada2 data into phyloseq at the end of the tutorial, so we could just integrate an example microbiome-package analysis based on that phyloseq object, with links back to your homepage/tutorial for more information.

[antagomir]

How about an example of selecting core microbiota based on user defined detection and prevalence cutoffs; and subsequent visualizations of the core? The "core" function should be intuitive and useful, therefore I was thinking of that as an example (we can look at other options, though).

Some examples are here, and the Rmd code.

Could you pick up something useful from this one already, and add to DADA2 workflow?

Or if you would like us to work through a small example, shall we use the example data from here: https://benjjneb.github.io/dada2/tutorial.html?

[benjjneb]

That seems like a good potential exemplar of the microbiome package.

I do want to use the data already in the tutorial if at all possible. It's not ideal (1 mouse "subject") but still should do as an example dataset I think.

[antagomir]

Perfect. Can you add the example based on the current material that we have in the core Rmd file that I linked above (it essentially has all the code that is necessary for a short example), or do you need help with this?

The DADA2 tutorial could link to microbiome tutorial (note the updated link).

[antagomir]

@benjjneb I noticed we somehow forgot this discussion. If this is still topical I could make a PR?

[benjjneb]

@antagomir

This has gone a bit stale, but it is still something I'd like to do. Especially given that phyloseq has become something like abandonware.

[antagomir]

PR done