Entering edit mode
9.5 years ago
marina-orlova
▴
90
Hi
I have results of two knockdown experiments: log2 values of expression changes of all genes after depletion of two different regulator proteins. How can I find out if these results are correlated? I guess I should not just count correlation coefficient but make some positive or negative control.
It sounds like you need a negative control of some kind, yes.
Is it an RNA-Seq experiment? If yes than a "chip-seq" tag is quite misleading.
Second, how could you have a log2 value of expression change without having a control?
And I think you could just correlate the variables, if these are fold change values. If your knockdown doesn't change anything, than log fold should be just ~normally distributed around 0, so this will cancel the correlation.
Thank you for correction, it's rna-seq
Control was present at both experiment, as for my question I mean control for comparison - to determine if correlation coefficient is high or low for such an experiment.
Ok, thank you for your answer, it helped.
Sorry, but you are going to have to be much more specific. You have log2 fold changes? How many samples? How are the samples related to each other? What two vectors of numbers are you planning to correlate?
I have log2 fold changes of genes' expression for two different knockdown experiments, about 350 samples (they are genes of drosophila m.)