I have made a UMAP of malignant cells and the result is split into 3 clusters, as seen below.
For the sake of example, let's say I want to find the differentially expressed genes in the uppermost cluster (the brown, red, and gray one). Does this methodology sound correct:
- Load in the EBSeq library. Use the EBTest function - for the Data argument, use the expression matrix for malignant cells. For the Sizes argument, run MedianNorm on said expression matrix. For the condition argument, create a vector denoting whether the malignant cells in the expression matrix are in the uppermost cluster or not (two conditions), for the maxround argument use a number that fits within that works based on convergence.
Are there any glaring flaws with this method? Am I using condition incorrectly? Should I be using all of the malignant cells as my input data when I am trying to find differentially expressed genes in the uppermost cluster?