Hello all,
I am new to using Minfi, and I am running into some issues with the estimateCellCounts function. I have looked through some forums but can't seem to find anything that addresses the issue. I am processing around 180 samples from EPIC arrays and am getting almost all samples having a composition value of 0 for the granulocyte estimation. Some things I have thought to look into are converting my EPIC rgSet to a 450k rgSet, filtering down to less CpG probes (around 30,000), and using FlowSorted.Blood.EPIC. Each of these is either giving me the same results (0's for most of the granulocyte values) or is not running.
Have you seen this before? Any ideas what might be the issue with this?
Thanks for your help!
I understand that you have whole blood samples and that's why you expect them to have more granulocytes? You could also try the ENmix cell estimation function, which is a bit more versatile with the input data and offers more reference datasets, to see if you get similar results (I seem to remember having problems with EPIC arrays in minfi cell estimation in the past).
Thanks for the reply! I tried the ENmix cell estimation function and it was much easier to run, but returned the same values for the neutrophils/granulocytes. Any thoughts?
Thanks again for all the help!!
If the functions output the same results, I would think the problem is with the input data/arrays. If the quality control (bisulfite probes, intensities, detection p-values, etc.) of the data is OK, it is difficult to say. Maybe the samples are not whole blood, but rather PBMCs (i.e. mononuclear white cells excluding granulocytes)? it is a quite common purification done during blood DNA extraction. Are you totally sure that the blood contained granulocytes? And what are the estimated counts? Mostly lymphocytes?
Haha, totally my fault! Just realized that they are PBMC samples, not whole blood. Silly mistake, thanks for the help!