Using paired data, I am comparing the lung microbiome with the mouth microbiome in 33 patients I used deseq2 just fine with this, however when I did it in 8 patients comparing the lung microbiome over two time points (again paired data) I got this error:

log2 fold change (MAP): time later vs baseline 
Wald test p-value: time later vs baseline 
DataFrame with 0 rows and 6 columns

Can anyone tell me why this would be occurring?

Thanks!

> library("phyloseq")

> packageVersion("phyloseq")

> otuinter=import_biom('C:\\Users\\closed-ref-32964076\\otu_table.biom')

> otuinter

> mapinter=read.csv('C:\\Users\\Qiime Maps\\R_map_intersputums.csv',header=T,row.name=1,stringsAsFactors=F)

> mapinter    

> sampledatainter=sample_data(mapinter)

> otumergeinter=merge_phyloseq(otuinter,sampledatainter)

> otumergeinter

> source("https://bioconductor.org/biocLite.R")

> biocLite("DESeq2")

> library("DESeq2")

> datainter=phyloseq_to_deseq2(otumergeinter,~Pairedinter+time)

resultsinter=DESeq(datainter,test="Wald",fitType="parametric")

resinter=results(resultsinter,cooksCutoff=F)

> alpha=0.05

> resultsinter2=resinter[which(resinter$padj<alpha),]

> resultsinter2

I think you are requesting a 0-row subset of the result.

I'm guessing that there are no adjusted p-values less than alpha. This could be because there are no differences, or the experiment and analysis is underpowered to detect the true differences.

summary(resinter$padj < alpha)

You can sort the results table by padj and see what the smallest value is (see DESeq2 vignette).

Thank you!  That really helped.

I got:

Mode   FALSE    NA's 
logical    1122       0 

So I'm guessing that means there are no differences.