If I have phyloseq file instead of example data how I show start

[Hesham999666]

Hi Guys So I am new to lefse analysis, I start using lefser I am starting with phyloseq file then I produced otutable and with a metadata file I made S4 object of SummarizedExperiment using the following code: counts = otu_table(phytted) %>%as.data.frame() colData = meta(physeq) %>% as.data.frame

phyExp = SummarizedExperiment(assays = list(counts = counts),colData = colData)

usually I got error that

Error in validObject(.Object) : invalid class “SummarizedExperiment” object: nb of cols in 'assay' (417) must equal nb of rows in 'colData' (409) NB: the extra columns is the taxonomy and OTU names in the tables
Even if i am able to fix that later on when I am run the lefser I got another error res.lefse.1 = lefser(phyExp.4, groupCol = "SampleType") Error in cols * 0.05 : non-numeric argument to binary operator

Is there is smooth way to lefser from phyloseq object or do you see problem in my code can be fixed

Thanks Hesham

[AbbiHern]

Did you ever find a solution to this? I have the same problem.

[christophevandijck]

I think you can use the function makeTreeSummarizedExperimentFromPhyloseq() from the mia package

[Hesham999666]

Hi, I ended using another package (microbial) they also have lefse analysis called "ldamarker"

ldamarker(physeq, group, pvalue = 0.05, normalize = TRUE, method = "relative")

Here is the code if you are interested.

upload the libraries

library(microbial) library(microbiome) library(ggpubr)

remove low abundance otu, threshold is 0.005% (0.00005)

otu <- otu_table(phyloseq) high_otu_idx <- colSums(otu)/sum(otu) >= 0.00005 high_otu <- taxa_names(phyloseq)[high_otu_idx] ps.lefse <- prune_taxa(high_otu, phyloseq)

perform the lefese on phyloseq

res <- ldamarker(ps.lefse ,group="SampleType")

lefse_plot <- plotLDA(res,group=c("X","Y"), lda=5, pvalue=0.05, color = c("red","green"), fontsize.x = 7,fontsize.y = 8)

Then I stopped using lefse and started to use ANCOM BC instead (https://www.nature.com/articles/s41467-020-17041-7) Analysis of compositions of microbiomes with bias correction. based on co-author advice more for compositional data like 16 sRNA bacterial communities.

• Hesham

[christophevandijck]

Dear Hesham, thanks a lot for your comment and suggestions! I will have a look at ANCOM BC, too.

[pauGuas]

@Hesham999666 I tried this code and my plot is coming up blank. Do you know what this could be?