I have a dataset of RNA-seq samples for testing different drugs on the presence of another drug. One of my samples is the normal cells with no drugs (control negative) and another is the cells with just the base drug (control positive). All of the other samples are cells with the base drug and another testing drug. So, my question is how to do design a proper Design Matrix for comparing these samples. I can separately compare with control negative and control positive. But, my intention is to somehow neutralize the effect of one of the controls in all the samples and compare to the other one. Is that possible? How should I design the matrix?

  • $\begingroup$ Literally one undrugged control? $\endgroup$
    – swbarnes2
    Commented Aug 2, 2020 at 17:18
  • $\begingroup$ Yes, Literally! Unfortunately I don't have biological replicates for any of these samples. $\endgroup$ Commented Aug 2, 2020 at 17:29
  • $\begingroup$ That's unfortunate. I guess you can do some ratios in excel, but you can't do anything sophisticated without replicates. And your conclusions will be really shaky. Wet lab people are going to have to test your conclusions with better experiments. $\endgroup$
    – swbarnes2
    Commented Aug 2, 2020 at 17:37
  • $\begingroup$ I know that. But the whole experiment's purpose was to generate some hypotheses for further testing with replicates. So, I guess it can give me some insights! $\endgroup$ Commented Aug 3, 2020 at 9:09
  • $\begingroup$ Or it might give who a whole lot of false positives that waste time and money. Everything hinges on your one replicate being perfectly representative of what the control should look like. $\endgroup$
    – swbarnes2
    Commented Aug 3, 2020 at 16:02


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.