Suggestions on papers for use as template when writing up QIIME 2.0-analyzed data

What recently-published paper(s) might you recommend using as a solid example of a complete and well-written work when trying to figure out how to write up QIIME 2.0-analyzed data? Any recommendations of such papers using paired-end sequences for a microbial diversity study would be a plus!

I am finding a great diversity of approaches that range from the very skimpy (“We used QIIME 2.0 to analyze the data”), to ones that go into so much detail I expect to see listed what the writer had eaten for breakfast and the color of his lucky bowling shirt… I am hoping to find at least one example that the community feels is a shining pillar by which all QIIME 2.0 papers should be judged.

2 Likes

Good question @rogergold!

I will let the community speak to the shining examples — but I just want to point out one resource that can help guide you and others, that gives a very minimal example of a partial methods section, which I hope may give my recommended level of detail on just a portion of the analyses:
https://docs.qiime2.org/2019.7/citation/

So clearly I’d tend toward the lucky bowling shirt description, having hung my head in sorrow far too many times at seeing papers that leave off at “We used QIIME 2.0 to analyze the data”, full stop. :disappointed: :sob:

Happy QIIMEing! :bowling:

4 Likes

I think researchers should attach a git repository with data and notebooks used with Qiime2 commands for reproducibility of their results.
I saw a lot of papers with bad description from which it is completely unclear what they did to obtain some results.

4 Likes

I love the people who publish a git pipeline (although god knows that going from what I view as good enough analysis code to good enough publication code takes forever, its like writing a second paper for me). But, sometimes I wonder about the utility in the absence of data accessability.

And, as far as bowling shirt and breakfast, (1) citing tools increases the probability that the tool developers can continue getting money to maintain their tools so its good for everyone if they can be cited; (2) reproducibility relies on being able to intuit the commands that were run. Its an issue everywhere, but maybe microbiome analysis can stand as an example of places where things made sense; and (3) have you not ever wanted to acknowledge/link to a playlist behind a paper?

5 Likes

Hello Roger,

Given that Qiime 2 has great tutorials based on real papers, you could check out how these paper were written up!
Fecal microbiota transplant: tutorial, method section
Atacama soil microbiome: tutorial, method section
Parkinson’s Mouse Tutorial: tutorial, method section

That a good companion to Nick's gold standard recommendation :1st_place_medal: and what's published by journals :2nd_place_medal:.

Note they all have something in common; when a method is used with defaults, 'exactly how I thought they did it,' the details are brief. When a method is unusual or defaults are changed, then more detail is given.


I think part of this depends on the journal format. A Nature Letter is only about 4 pages long, so keeping methods concise is paramount. And they only allow ~30 references, and you could easily cite 30 pieces of software.

When brevity is demanded, linking to a computational notebook on GitHub or posting your .qzv file with full provenance is a good alternative to a detailed method section.

Colin

P.S.

I expect to see listed what the writer had eaten for breakfast

Geese, what's wrong with a detailed method section? Were you making a joke?

3 Likes

Just an addendum: these may give you a sense of style, but not of the methods themselves, since those articles used QIIME 1, and the methodology has in many cases changed considerably.

3 Likes

Excellent suggestion, thank you! I may have some more questions about this later on!

1 Like

Thank you all for the excellent suggestions! I sincerely appreciate your great advice. I may have additional questions later on, but these resources will help a lot to get me started! Thank you!

1 Like