Hi,

I have datasets for 3 different conditions["WT", "KO", "KO1"] (each with 3 replicates). [and two time points["0" and "12"]].

To use in DESeq, I have constructed a DSElist, with design ~ group where group has 6 levels.

dds <- DESeqDataSetFromMatrix(countData = readCount,
                                 colData = readCountColData,
                                 design = ~ group)
dds <- DESeq(dds)

My objective to find genes that have higher expression in KO compared to WT and have higher expression in WT compared to KO1. This all at a single time point, lets say "0". Would it be correct if I get the difference in gene expression between KO vs WT = res1 (select log2FoldChange > 1.5) and WT vs KO = res2 (select log2FoldChange > 1.5). And then I try to find which genes in res1 have log2FoldChange higher than 1.5 compared to res2. Is this the correct way?

res1 <- results(dds, contrast = c("condition", "KO", "WT"),
                 alpha = 0.05)
res1 = na.omit(res1)
res1 = res1[res1$log2FoldChange < 1.5,]`

res2 <- results(dds, contrast = c("condition", "WT", "KO1"),
                 alpha = 0.05)
res2 = na.omit(res2)
res2 = res2[res2$log2FoldChange < 1.5,]

genelist = intersect(rownames(res1), rownames(res2)
genelist = genelist[genelist$log2FoldChange < 1.5]

Yes I would approach this by just looking at the intersection of the two FDR sets.