Hello I'm trying to interpret some results and to find a rational to explain what I see from the numbers I obtained. I suppose that this is a common possibile outcome and I'm having problems in explaining and justify with words what I see. In my specific case, these are the results from Diffbind analysis after chipseq on H3K4me3 but it can happen also for other types of analyses, such as differential analysis in rna-seq.
I have three conditions : A, B and C.
A = 3 months old
B and C = 12 months old
B= diet A
C= diet B
A vs B: 775 differences
A vs C: 743 differences
If I intersect these sites, 636 are unique in one of the two comparisons, while the others are present in both. So if I have just these results I'm brought to say that the 636 unique sites are dependent from the diet or at least the trend is significant just for one or the other diet.
But if I directly compare B to C, I find just one difference (and in this case is not in the 636 unique sites but it is present both in A vs B and A vs C).
The site found in the comparison between B and C is the only that reaches statistical significant but is it true that the is something going on also in the 636 sites??And how it can be reported from a statistical point of view?
I hope I was sufficiently clear
Thank a lot
Francesca
