Hi, I’m using LEfSe 1.1.01 on Linux (Ubuntu).
I have five groups in my data, and I want the significantly abundant taxa to be identified for each group.
I’ve noticed that the results differ depending on the number of groups included in the input data. For example, the results are different when the input data consists of five groups compared to when it consists of only two groups.
lefse_format_input.py genus_data_3w.txt genus_data_3w.in -c 1 -u 2 -o 1000000
lefse_run.py genus_data_3w.in genus_data_3w.res -y 0
As you can see in the code, the analysis was done in one-against-all setting (less strict), so it should not cause the multiclass problem.
I assume that the results differ in cases where a taxon is significantly different in multiple groups, but only the group with the highest LDA score is reported in the .res file.
Can anyone please tell me how to report all the groups that satisfied the p-value cutoff (not just the group with highest LDA score) in the .res file, and in the LDA score bar plot?
Thanks in advance.
Jihyun