Lefse command skipping all iters because too few groups or not enough contrast

[jmacnico]

Greetings,

Experiment overview:
I have 12 subjects. Each of my subjects had a treated and an untreated sample taken (A and B) over 5 time points (0, 3, 7, 14, 30). I also have an additional control group from a skin scraping for 2 of my subjects ( C ).

Command:
I am trying to run a lefse command on my data and all the iters error out in this way:

[WARNING]: Skipping iter 1 in LDA test. This can be caused by too few groups per class or not enough contrast within the classes.

I am using the command: mothur >
lefse(shared=stability.trim.contigs.good.unique.good.filter.unique.precluster.pick.pds.wang.pick.tx.1.subsample.1.pick.shared, design=status1.design)

My status1.design file is set up as:
Sample. Group
1_A_0. A0
1_A_3 A3
ect
1_B_0 B0
1_B_3 B3
ect

I am wondering if one of the options for the lefse command might help with this?? I checked the command on the MothurSOP but was unsure which option might best help.

Many thanks. JM

[pschloss]

It sounds to me like lefse is not the command you want. You likely want a paired comparison (treated and untreated from same person) and repeated measures (same person over time). Lefse won’t do that. I’d suggest talking to a statistician about the proper test for this design.

Pat