Seurat's DESeq2 option on FindMarkers() returns different results from standalone default DESeq2 pipeline

[CamiArcu]

Dear Seurat Team;

I am having troubles in finding how Seurat handles DESeq2 in differential expression analysis. I have been looking to differential_expression.R but I couldn't find the function PerformDE() to get the parameters used by DESeq2 inside FindMarkers().

I supposed in this scenario that the default DESeq() function in used to perform the analysis, but I have tried:

data <- DietSeurat(SeuData)
data  <- as.SingleCellExperiment(data)
data  <- data[features, ] # select specific features according to a pre-filtering step
assay(data) <- as.matrix(assay(data))
data <- scran::convertTo(data, type = "DESeq2")
design(data) <- formula(~Treat)
data <- DESeq(data)
data_res <- DESeq2::results(object = data)

And ordering by pvalue, I get a different rank than doing:

SeuData_res <- FindMarkers(SeuData, features = features, min.pct = 0, assay = "RNA", slot = "counts", ident.1 = "PD", test.use = "DESeq2", logfc.threshold = 0)

and ordering by p_val.

So I am just wondering if you can take me to the link where I can found the source code of PerformDE() or maybe you can correct my script above to know how DESeq2 is handled.

Thanks a lot for all your work in the package and your help through the discussions.

Best,
Camila