Downsampling or not in single cell RNA Seq Experiments

Question

[MolGeek]

Hallo everyone,

New to scRNA data analysis here. I have 10X data for 2 conditions (2 replicates each) as WT and Sickle and i want at first to check for visual differences between the clusters provided by UMAP using Seurat5. All the experiments have different number of cells.

Question n1: After Qc should i downsample the data in order to have even number of cells in each condition or not?

Question n2: If i have to downsample, should i merge the two replicates of each condition and downsample by condition or should i downsample to an even number all replicates and then proceed to integration?

Thanks in advance.