LEfSe without taxonomic hierarchy - still valid?

Downstream analysis and statistics LEfSe
1

Hello,

Currently I have two questions about running lefse

  1. I didn’t use the subclass option (i.e. my command looks like format_lefse.py ~.txt -c 1 -u 2 -o 1000000)

  2. Most importantly, I didn’t use the taxonomic hierarchy structure.

The purpose of analysis is just looking for specific taxa level marker(e.g. genus level) in 2 groups (e.g. CRC/healthy)

In this case, above two are still making lefse analysis valid??

Thank you.

2

Hi -
In your case you’re essentially running a Kruskal-Wallis test on the features. It’s still valid for all the features tested, but just bear in mind LEfSe’s additional functionality (filtering by effect consistency among subclasses for example) are not activated.
Best,
Siyuan

1 Like