I have a question regarding the p and q values that I observe in my statistics of analysing protein expression.
I get the following p values that are very significant, however the Q values are very high. I have performed this analysis in a Software called progenesis and was hoping to analyse this data in manners described in this forum.
Please see the image attached. What do you conclude from this output - can I not trust the p-value because the q value is so high? Is my study underpowered and perhaps I need a higher n number?
There is a description of what the q and p value is, however it is not clear if they have adjusted the p value according to the multiple testing or not.
Link here: http://www.nonlinear.com/support/progenesis/comet/faq/v2.0/pq-values.aspx
We would like to proceed with a bigger experiment based on this data, however we are not confident whether the effects seen are due to the random testing issue, study being underpowered or any other issues.
Thanks for the answer. We came to the conclusion that the sample size was too low. While we see an effect, the noise in between samples introduces much more variability. When performing PCA analysis we can see an effect with the treatment however some samples are close to the controls: hence the FDR cut off ratio is high making the Q values to be uncertain. So the decision was to increase the sample size.