Possible to calculate diff_analysis with less than 5 sample per group ?

[UpalabdhaD]

Hi , I was interesented in calculating diff analysis. My dataset include 6 sample, 3 per group.

So i was wondering is it possible to run diff analysis for less than 5 sample per group ?

• When I run diff_analysis:

  diff_analysis(obj = rf_pseq_obj, classgroup = "Status",
                     mlfun = "lda",
                     filtermod = "pvalue",
                     firstcomfun = "kruskal_test",
                     firstalpha = 0.05,
                     strictmod = TRUE,
                     secondcomfun = "wilcox_test",
                     subclmin = 3,
                     subclwilc = TRUE,
                     secondalpha = 0.01,
                     lda=3)

• got the following error:

Error: The sample number of each groups is smaller than 5.
Traceback:

1. diff_analysis(obj = rf_pseq_obj, classgroup = "Status", mlfun = "lda", 
 .     filtermod = "pvalue", firstcomfun = "kruskal_test", firstalpha = 0.05, 
 .     strictmod = TRUE, secondcomfun = "wilcox_test", subclmin = 3, 
 .     subclwilc = TRUE, secondalpha = 0.01, lda = 3)
2. diff_analysis.phyloseq(obj = rf_pseq_obj, classgroup = "Status", 
 .     mlfun = "lda", filtermod = "pvalue", firstcomfun = "kruskal_test", 
 .     firstalpha = 0.05, strictmod = TRUE, secondcomfun = "wilcox_test", 
 .     subclmin = 3, subclwilc = TRUE, secondalpha = 0.01, lda = 3)
3. diff_analysis.data.frame(obj = otuda, sampleda = sampleda, taxda = taxda, 
 .     ...)
4. diffclass(datasample = datameta, features = varsfirst, comclass = compareclass, 
 .     classgroup = classgroup, fcfun = fcfun, secondcomfun = secondcomfun, 
 .     classmin = clmin, clwilc = clwilc, pfold = secondalpha, ...)
5. get_gfc_wilc(datasample = datatmp, fun1 = fcfun, classlevelsnum = clsize, 
 .     vars = features, classname = classgroup, minnum = classmin, 
 .     fun2 = secondcomfun, wilc = clwilc, ...)
6. stop_wrap("The sample number of each groups is smaller than ", 
 .     minnum, ".")
7. stop(msg, call. = FALSE)

Thanks in advance

[jianlingfan]

I have the same probelm. Have you solved it?

[xiangpin]

please check whether each group's sample number is smaller than the clmin argument. PS: We suggest using mp_diff_analysis to replace diff_analysis, you can refer to the vignettes