Hi! I am doing pairwise comparisons with DESeq2 (version 1.24.0). I would like to shrinkage the log2 Fold Change using the normal approach. I used the following code:
dds <- DESeqDataSet(se_sel, ~ condition) dds <- DESeq(dds) resNorm <- lfcShrink(dds, contrast=c("condition", cond1 , cond2 ), type="normal") resNorm log2 fold change (MLE): condition cond1 vs cond2 Wald test p-value: condition cond1 vs cond2 baseMean log2FoldChange lfcSE <numeric> <numeric> <numeric> ENSG00000015592 2252.24746949959 4.93239360176146 0.129430249903588 ENSG00000083857 38573.0261803919 3.96140286121604 0.0734986203310571 ENSG00000090339 2854.89111431782 4.08839129938656 0.10107762978087 ENSG00000171551 12450.6367006511 6.65073841803356 0.132315837227065 ENSG00000076356 8747.08707079265 -3.53308614922331 0.0955398226626009
I was wondering if it is correct to shinkage the LFC with the function lfcShrink or considering that I am using normal approach it is better to run DESeq function with betaPrior=TRUE and then use the function result with contrast to extract the results. In addition, I will considered in my analysis the file with shunkage LFC (resNorm). Is that correct?
Thank you for your help!