I agree that this is a huge dataset and very demanding of compute resources!
That said, we’ve tried to set up the Transcriptomics Explorer to support common forms of investigation. For example, you can add one or many genes to the heatmap, and you can visualize the expression for a single gene across the entire dataset in the Scatterplot view.
When you say “filter for the genes I am interested in,” what kind of operation are you trying to achieve?