Dear all, I have a question regarding the way I am interpreting my results. I really hope you can give me a hand. I have four groups that I am analyzing: 1) WT - vehicle 2) WT - treatment 3) Mut - vehicle 4) Mut - treatment
I am trying to get a list of DE genes, affected by the treatment. As a first step, I ran DESeq2, following the official vignette:
dds <- DESeqDataSetFromTximport(txi, colData = coldata, design = ~ genotype + condition) #set levelels dds$condition <- relevel(dds$condition, ref = "vehicle") dds$genotype <- relevel(dds$genotype, ref = "WT") #I run the deseq function dds <- DESeq(dds) #and then I got the results using "condition Treat vs vehicle", to get the comparison between treated and untreated res <- results(dds, name = "condition_Treat_vs_vehicle") # and then I got the results using "genotype_Mut_vs_WT", to get the comparison between treated and untreated res2 <- results(dds, name="genotype_Mut_vs_WT")
In order to understand which genes are affected only by treatment without any genotype effect, I then ran the LRT test.
dds_lrt <- DESeq(dds, test="LRT", reduced = ~genotype) res_LRT <- results(dds_lrt)
Then I selected the significant genes via
padj <0.05 for
And I overlapped these lists of genes using a Venn Diagram function to see how many were actually affected only from my treatment.
Now, I doubt that this comparison is actually not correct because I should only use the results from my LRT test, without considering the ones obtained via the Wald test.
It would be great to get more insights on this topic. Please let me know if there is any information missing that might help the answer.
Looking forward to your feedback
Thank you in advance :)