Interpreting QIIME2 Outputs: Dos and Don'ts

Hello team,
I believe that many of the qiime2 users in this forum have published research papers that provide results based on the outputs of qiime2 pipeline. QIIME2 has certainly accelerated the pace of research in reverse ecology. However, it is important that in the race to publish, we do not overlook, ignore, misuse, misinterpret or violate the fundamentals of sequence analyses. By doing so, we not only contribute to pseudoscience but also overdo science claims. As an Editor/Reviewer of journals in this field, I come across a large number of manuscripts and also read several published articles that violate/have violated the assumptions/theories of reverse ecology and sequence analyses.

To minimize pseudoscience, we must form a team consisting of subject experts, computational biologists and data scientists to review articles that used QIIME2 pipeline but violated the assumptions underlying each q2-plugin. Publishing a commentary on the erroneous article would caution other researchers against the overuse of q2-results. Unfortunately, not all journal Editors/Reviewers who handle such erroneous manuscripts are experts in QIIME2, so they fail to detect the violations.

I urge the q2-admins and users to come forward and provide their ideas to this proposed initiative. Thanks!

BS

6 Likes

Hi @bsen2018 ,

Well said!

I think that a commentary/review article on common pitfalls and test assumptions would be a novel and useful article! This is something that reviewers and editors can use as a checklist to ensure appropriate use of sequencing data.

However, a few ideas to add:

These issues/pitfalls are not specific to QIIME 2, but to all sequence analysis. So addressing QIIME 2 outputs specifically would be too narrow. The violated methods are not unique to QIIME 2 but are commonly used in the field.

Thousands of articles cite QIIME each year! So the volume is quite high and it would take a large team to review. Moreover, the burden should be on the journals/editors/reviewers to validate the quality of individual papers — but I see your point that beginning to write such commentaries would pressure authors/journals/editors/reviewers to do their due diligence.

Moreover, I think that a "name and shame" approach is not productive except in extreme circumstances, since it is a somewhat toxic and public way to confront the issue. I think that education (e.g., community documentation, review article(s), checklists, etc) is a more proactive and supportive way to spread awareness.

9 Likes

When I was learning the skills of qPCR, I found a website called MIQE: minimum information for publication of quantitative real-time PCR experiments.I think they draft that guidelines in a conference. Maybe some of the experts in this forum can guide us to do similar thing.

1 Like