how to show highly expressed genes between samples with the same condition

Question

[minoo]

I have three samples of RNAseq showing count values of gene expression which all of them are control conditions. Two of these samples are from same patient and the third sample is from another patient. I need to show the highly expressed genes within all the samples. To this end i used Deseq package like below:

dds <- DESeqDataSetFromMatrix(countData = cstm, colData = colData, design = ~ 1)

but the result shows the genes that are highly expressed in one sample which belonged to different patient in comparison to three samples. However I didn't enter the patients info in the coldata as I didn't want to be considered as a factor in this analysis. Does anyone a=has a better understanf=ding of how to show highly expressed genes within samples of similar conditions?

Answer 1

[shiyang_bio]

If I understand correctly you want to show the genes with high expression level in all 3 samples. If so, Deseq2 is not needed. You can just calculate the mean expression of each gene and select those highest ones. Or you can set a cutoff, say 10 (in FPKM or TPM), and select only genes with > 50 FPKM/TPM in all samples.

[cpad0112]

 You can just calculate the mean expression of each gene and select those highest ones

Sample variance and gene variance needs to be adjusted. Simple mean would not account for gene/entity expression variance across the samples and difference among the samples. Values must be adjusted and then OP can take the entities of high expression.