Hi!

I tried to see how NB model is suitable for microbiome data as mentioned in DESeq2, I used this junk of codes. Could you correct me please, if I am out of the track? I used phyloseq object 'dietswap'.

library(DESeq2)
library(microbiome)
library(ggplot2)
data("dietswap")
a <- rowMeans(abundances(dietswap))
dfs <- data.frame(taxa = names(a), meanabundance = a) %>% arrange(meanabundance) %>% mutate(taxa = factor(taxa, levels = rev(unique(taxa))))
p <- ggplot(dfs, aes(x = taxa, y = meanabundance)) +
geomhistogram(stat = "identity") +
labs(x = "taxa", y = "Mean read count (N)", title = "abundance curve")
print(p)

I don't recommend DESeq2 for microbiome data, as I did no testing on microbiome data during the development of the software, nor am I an expert in this area of bioinformatics.