Hi, A lab member has the results cvs file from DESEQ2. I am trying to recreate the results given that she lost her initial runs as her hard drive crashed. However, I get many less significant genes (100s). I think I am using the same counts and procedure as she used. The results have the same base mean for the overlapping sets. However the log2fold difference are different. I feel the dispersion calculations must be distinct between what I am running and what she has run. Has this changed between versions? Maybe I am using different parameters then her? I have not tried different modification parameters yet to try to see if I can get back her results? Can someone suggest parameters/techniques or what to look at to diagnose this problem?
With a different version of DESeq2 it's perfectly possible that you get different results. And if you aren't sure about the parameters which were used earlier, then it's even more doubtful that you can replicate the results.