Bulk-RNA-Seq integrate control samples.

Question

[fabian]

Hi,

I have Bulk-RNA seq data from patients who had a heart attack, including two groups: those who went into shock and those who did not. Additionally, the dataset includes transcriptomes from 5 healthy individuals who did not have a heart attack.

I want to compare the transcriptomic signatures of patients who went into shock after the heart attack versus those who did not. My assumption is to use the heart attack patients without shock as the control group and exclude the healthy samples, as they are unrelated to my primary question.

However, I am new to biostatistics and wonder if this approach is correct or if there’s a way to include the healthy samples to add more depth to the analysis.

My model matrix would look something like this:

design <- model.matrix( ~ 0 + shock, data = targets)
contrasts <- makeContrasts(
  "Comparison" = YesSchock - NoShock, levels = design)

I am using R and the "From Reads to Genes" Pipeline for my analysis.

Answer 1

[i.sudbery]

No, I'm not aware of any other way to include data from the healthy controls if your research question is "How does shock effect the gene expression when having a heart attack".

If, for example, you compared each sample to the health controls, and then compared the differences to each other, you'd simply end up with the same comparison:

(YesShock-Health) - (NoShock - Healthy) = YesShock - Healthy - NoShock + Healthy = YesShock - NoShock

If you wanted to know how going into shock affected the gene expression changes caused by having a heart attack, then you'd need before and after heart attack data for each patient, which, obviously, would be more or less impossible to obtain.