I have a proteomics data from MaxQuant software. According to the manual, the data has been normalized by the MaxQuant internal method. After cleanup, QC and imputation (imputation data), I have done the adjustment with limma for AOD, Sex and PMI (residual data). In the next step, I want to plot a DE heatmap for case and control. Which data should I use, the imputation data or residual data?