Clade error in LEfSe - clade_sep parameter too large, lowered to 0.200225830078

Downstream analysis and statistics LEfSe
1

Hi all,

I am trying to run LEfSe on Galaxy using a merged Metaphlan3 output. However, when I want to plot the cladogram I have the following figure and the message “clade_sep parameter too large, lowered to 0.200225830078”.merged_all_0503_metaphlan_abundance_table_reformatted.txt (11.7 KB) .

Galaxy69-D)_Plot_Cladogram_on_data_67
Galaxy69-D)_Plot_Cladogram_on_data_671200×900 49.3 KB

Could you please help me to figure out what is the problem?

Thank you very much in advance for your valuble time and concern!

Best,

Daniel Davila

2

Hi Daniel,
Thanks for your question. Looking at your data set, it seems like you have only one observation per group–is that correct? I think that’s the reason you’re seeing unusual results. Let me know if I’m misunderstanding!
Best,
Meg

3

Hi Meg,

Thanks a lot for the time and help! I have only one observation per group on this table because I would like to figure out the difference between observations in different groups. However, I have more observations from the same groups. If I understood correctly, I should perform the analysis with all the observations for each group. Is it correct?

Thanks again for your time and concern!

Best,
Daniel

4

Hi Daniel,
Yes, you should include all observations across all groups, and include a row that specifies which group each observation belongs to–this variable should be specified as the “class” variable when running LEfSe.
Please let me know if you have any other questions!
Best,
Meg

5

Hi Meg,

Thanks for your valuable help! I tried again with your recommendation an I still have the same error. This is the new data set that I am using MAM_merged_abundance_table_reformatted.txt (21.5 KB). and the figure that I get is the following

Galaxy92-D)_Plot_Cladogram_on_data_91
Galaxy92-D)_Plot_Cladogram_on_data_911200×900 65.9 KB

What do you suggest? Thanks a lot for your help and valuable time!!

Best,

Daniel

6

Hi Daniel,
It looks like there’s no features which are differently abundant between the groups in your class variable, which is why none are showing up in the cladogram. When you run part B) LDA Effect Size on the Galaxy interface, your results should look something like:
“Number of significantly discriminative features: 0 ( 0 ) before internal wilcoxon
No features with significant differences between the two classes
Number of discriminative features with abs LDA score > 2.0 : 0”
From that, you know that plotting the features or cladogram won’t be useful.
Let me know if you have any other questions!
Best,
Meg