How to perform LEFSe after qiime2

[ehdud0505 (Doyoung Jin)]

Hi, everyone. I am using qiime2-2024.2 version.

I want to run LEFSe and Picrust after 16s rRNA amplicon sequencing analysis, but it seems that qiime2 doesn't support it anymore.

So I'm writing to ask how I can perform both LEFSe and Picrust.

I checked that I can do LEFSe via R, but that also doesn't work well due to artifact issues...

I don't even know which data table I should use, so I would appreciate your help.

Looking forward to your reply, thank you.

-Doyoung-

[jwdebelius (Justine Debelius)]

Hi @ehdud0505,

For PICRUSt, there's a qiime2 plugin: QIIME 2 Library

Or, you can export your feature table and use -PICRUSt2 as a stand alone. Its worth considering wehther yoru enviroment is appropriate for functional prediction.

My advice is to strongly consider whether LefSe is appropriate for your work in the year of our microbes 2024 and I would generally say the way to perform LefSe after you build a feature table is to put your data through ANCOM-BC or Aldex2.This is a flippant answer, but here's my logic.

• Differential abundance analyses of relative abundance or rarefied data are inappropriate ( Weiss et al, 2017; Lin and Peddada, 2020 among others)
• That microbiome data is composititional, and this is not optional (an assumption LefSe doesn't make) (Gloor et al, 2017; Lin and Peddada, 2020)
• LefSe doesn't do FDR correction for multiple hypotheses (Segata et al, 2011; Nearing et al, 2022; )
• LefSe is inconsistent with other, more appropriate techniques (Nearing et al, 2022; )

Best,
Justine

[ehdud0505 (Doyoung Jin)]

Hello @jwdebelius,

Thank you so much for your kind reply, it was very helpful.

Thanks to the paper you mentioned, I was able to understand which analysis tool I should use.

I really appreciate it

Sincerely,

-Doyoung-