Hello everyone. I have obtained nCounter data on tumor samples of 4 distinct subgroups of one tumor type. I want to perform differential expression analysis between these 4 subgroups and get a heatmap that can differentiate between them.

My search tells me that because nCounter data are read counts (like in RNA-seq) I could use the DESeq2 package in R to do it. However, some of them also refer different types of normalization methods for the data. I am very new to bioinformatics analysis and have a hard time understanding these nuances. I have both the raw data and also the normalized data that was normalized in nSolver using their pre-defined method.

What shoul I be doing to perform the differential expression analysis? Use the raw counts data directly into DESeq2? Use the normalized data? Use a different package than DESeq2? Is it possible to perform this analysis between 4 different groups?

Hope you can help me and thank you in advance!

Please see the vignette which instructs to us raw counts. Nothing else will do for DESeq2. limma can either use raw counts with limma-voom or normalized data (typically logCPM-like) via limma-trend for count data. Again, please see its manual. There is also many previous questions on DESeq2/limma on Nanostring, please find them via a google search.