I use a cell line, in which, as all in vitro cell-based experiments, the passage number will affect the gene expression (the influence of the passage number is well-known and accepted in literature).

I have N=2 (which N=the biological replica that corresponds to passage number). Each N has 1) control, 2) substance A, 3) substance B, and 4) substance C (total number of samples=8, 4 for each passage number). In theory (literature-based), all substances (A, B, and C) are in the same category of drugs (to treat the same disease), and A and B are the same drugs, but B is slightly modified, and C is the positive control (previously have been tested in clinical trials and considered a gold standard in in vitro experiments).

The aim of the study is: 1) the differential expression between the control and each of the groups (A, B, and C). The comparison among the A, B, and C substances that will be represented in a heatmap (a further subcategory of custom genes that I will provide the GeneIDs of the categories- total three categories-), volcano plot and 2) the enrichment gene pathway for further indications in biological responses.

```r dds <- DESeqDataSetFromMatrix(countData = cts1256, colData = coldata1256, design = ~ passage + treatment ) dds <- DESeq(dds)

keep <- rowSums(counts(dds)) > 10 dds <- dds[keep,]

sizeFactors(dds)

res <- results(dds) #,contrast = (c("passage","1","2")

res

vsd <- vst(dds, blind = FALSE) rld <- rlog(dds, blind = FALSE)

Does anyone come across a similar problem?

This cannot generally be answered. From a statistical standpoint it does not matter because if you use the exact same cells for all treatments then any effect due to passage should be the same everywhere. From a biological point of view it is on you to determine whether cells are too old to use or whether your treatments affect cells differently depending on passage, but that's beyond the scope of this (or any) forum and needs to be found out by experiment.