DESEq2 results : low counts too many genes

Question

[shome]

My deseq2 results shows as follows :

out of 55357 with nonzero total read count
adjusted p-value < 0.1
LFC > 0 (up)       : 0, 0%
LFC < 0 (down)     : 12, 0.022%
outliers [1]       : 0, 0%
low counts [2]     : 50430, 91%
(mean count < 13)
[1] see 'cooksCutoff' argument of ?results
[2] see 'independentFiltering' argument of ?results

Around 91 percent of genes are being considered as low counts.I checked for errors or alignment issues;but there isn't any.There is a possibility of one gene being too highly expressed in some of samples though.How to address the issues to make sure I am able to catch differential gene expression properly.

I checked PCA plots and batch effects etc..the data is totally fine.

Answer 1

[ATpoint]

Low count just means that these were removed by the independent filtering. See the vignette and DESeq2 paper on details. Essentially, you have almost no DEGs, maybe because there are none or because the study is underpowered. Anyway, what DESeq2 tries in the IF is to maximize the number of genes below the alpha threshold. If you have low number of candidate DEGs then it filters a lot. That is what you see here. For more guidance you would need to add some plots, like PCA and plotMA output plus the sample size and design.