edgeR differential gene expression strange p-value distribution

Question

[lluc.cabus]

Hi,

I'm doing a differential gene expression between samples with two different conditions (with only 6 samples in each group) following the tutorial from edgeR, but I have seen that the p-values follow a strange distribution in the volcano plot. The samples with higher logFC have lower p-values and in general the volcano plots don't seem like the normal plots. Do you know why could this be happening or if this is normal?

The code that I run is this one:

dgList <- DGEList(counts=FL_count_matrix[,1:12], genes=rownames(FL_count_matrix), group = as.factor(FL_metadata_EDTA_SD$Tube_type))
dgList <- calcNormFactors(dgList, method="TMM")
dgList <- estimateCommonDisp(dgList, verbose = T)
dgList <- estimateGLMTrendedDisp(dgList, verbose = T)
dgList <- estimateGLMTagwiseDisp(dgList)
fit <- glmQLFit(dgList)
qlf <- glmQLFTest(fit, coef=2)

Thank you very much, Lluc

[ATpoint]

Didn't we discuss this already here? Volcano plot with FDR values looks weird How is the question different from the previous one? It looks "odd" because you have many large p-values and no or few DEGs, so the Volcanos are pretty much bunched together vertically rather than nicely separating on the y-axis.

[lluc.cabus]

Yes ATpoint , I understand that I have many large p-values and low logFC and in the end I find few differentially expressed genes, but my question was more in the lines of why do the points of the p-values seem to follow a kind of quadratic function (in the first plot). Normally in the volcano plots the p-values are more "randomly" distributed, but they seem to follow a pattern here and I don't know if this is normal. This is one of my first differential gene expression analysis, so I'm not very familiarized with this, sorry if this is too obvious