Hello,

I am a R&D engineer trainee in bioinformatics. I am currently working on RNA sequencing data on R.

I need to determine the influence of bacterial supernatants on genes differentially expressed by cells in the presence of a drug. The problem is that in an experiment, the values of the parameters studied in a condition must be related to the control, so here on DESeq2 the control is supposed to be the cells alone and not the cells + drug, right?

So I thought of an alternative approach: i) determine the differentially expressed genes between condition A (cells+drug) and a control (cells) ii) determine the genes differentially expressed between condition B (cells+drug+bacterial supernatant) and a control (cells) iii) remove the genes detected common to both conditions A and B and extract the genes that are only detected in condition B.

I hope, from these extracted genes, to obtain the biological pathways that occur when cells are in the presence of the drug and bacterial supernatant. What do you think about this? Is it possible to perform a downstream GSEA analysis with this approach, or can I still perform a direct comparison of cells+drug+supernatant vs cells+drug? What modifications and/or pipelines would you suggest to me otherwise? Thank you in advance for the time and energy you will put into my query.

There is a difference in differentially expressed genes depending on the method used... condition A (cells+drug) vs control (cells) -> 956 differentially expressed genes condition B (cells+drug+bacterial supernatant) vs control (cells) -> 1258 differentially expressed genes -> 680 different genes between condition A and B However, when running condition B vs condition A on DESeq2, only 127 genes are differentially expressed I can't explain the origin of this rather large difference...

Sincerely,

Aurélien Y.

Code should be placed in three backticks as shown below

# include your problematic code here with any corresponding output 
# please also include the results of running the following in an R session 

sessionInfo( )

I don't see anything wrong with focusing on a group of genes that are DE in one comparison but not in the other (this has been posted previously but is probably hard to search for). Note that, when you want genes _not_ DE, we recommend specifying a minimal, biologically relevant LFC to define as "not DE" and use altHypothesis="lessAbs" and lfcThreshold`.