Hello! I have some doubts about what would be the best approach to analyze a particular RNASEQ dataset, given the experimental design.
We have 2 time points, one before randomization (PRE) and another post randomization (END), to 3 experimental arms (ABC). The metadata looks something like this. All "PRE" are technically the same. We want to compare between the groups at the END, point.
dds <- DESeqDataSetFromMatrix(countData = round(data_fil),
colData = reorder_metadata,
design = ~ pca_group1)
Initial PCA analysis shows a big Sex difference regardless of treatment or timepoint, which I think is biologically relevant. But also it seems like there is more differences between donors than with treatments. I was wondering what was the best way of accounting or correcting for this.
If I include de variable "patients" as part of the DESEQ2 formula, the difference between sexes gets lost. As of now I am just using Michael's Love vignette (https://github.com/mikelove/preNivolumabOnNivolumab/blob/main/preNivolumabOnNivolumab.knit.md ) and using RUVSeq for correcting some of this interdonor effect. Is there any way of using the "PRE" to normalize for inter individual variance?, or would this not be appropriate? Any help would be greatly appreciated. Thanks a lot!!!
