If I have 2 different codesets from nanostring, how should I run "RUV_total" function?
"Different codesets" means the dimension of 2 or more sets is different, such as below. There are a set of Endogenous genes are in common between the 2 sets. The house keeping (hk), negative, and positive genes are same for 2 sets. I have 2 ways in my mind to run "RUV_total" function. But, I have no idea which one is more reasonable, or better ideas?
First, I can merge the codesets, and then run "RUV_total" function on the merged one. For example, there are 241 Endogenous + 20 (hk, neg, pos) genes for 80 samples after merging. In this way, the genes which are not in common were removed, my concern is that if the removal of uncommon genes will have effect on ruv normalization.
Second, I can do normalization on codeset1 and 2, respectively. Then, the RUV factors can be combined to be pData and passed to DESeq2. In this case, the "countData" is still same as the above way.
Not sure if the "codeset size" (like the library size for RNAseq) has effect on ruv normalization or counts(dds, normalized=TRUE) function.
Any ideas or suggestions would be appreciated!