Entering edit mode
7.0 years ago
Muhammad Haseeb
▴
50
Hi,
I hope you all are doing good. My question is that I have log2 gene expression data of four different time span (30min, 1hr, 2hr and 4hr). I want to calculate the Fold change (FC) and P-value for it, Can I calculate FC individually for each time. For example (30min_FC,1hr_FC, 2hr_FC and 4hr_FC)? The number of columns is (10X4=40), in that case, we have 40 different fold change w.r.t time and similarly P-value. What can I do in this case?
Thank you,
Fold change is always a condition compared to another. You could compare all timepoints against the first.
Dear Wounter, Thank you for the answer, Can you clarify me, I didn't understand what do you mean by all time points against the first? You mean all time point that is 40 columns of fold change for each time, compare with control? For example: (30min_fold change of normal Vs 30min_fold change of treated......and likewise for all 40 columns of treated)
Wouter was just going off the data structure you described. You didn't mention treatment or control conditions! As Wouter points out, FC is a pairwise comparison. For your treatment v Control that would mean (using your terms here): 30min_FC = 30min_normal / 30min_treated.
For time-series data, people often calculate FC across time, but again since its pairwise you need something to compare to, usually t=0. So for you it would be: 1hr_FC = 1hr / 30_min.; 2hr_FC = 2hr / 30_min etc. You can do all 12 comparisons if you'd like (30min to 1hr, 30min to 2hr, 30min to 4hr, 1hr to 30 min, 1hr to 2hr etc.) but not sure you'd want to. Additionally, It's probably more informative to do this calculation on the control series...