Goal: can we efficiently perform a single-cell analysis fully in memory, even for a big dataset like the 1.3M cells dataset? As long as the machine has enough memory.
Here are some low-hanging fruits that have the potential to significantly improve the current situation.