After the above quality control, all cells from the Seurat objects for each pool were integrated for visualization in the same 2D space. Atomic sketch integration was used, a dictionary learning based procedure for large datasets (see https://satijalab.org/seurat/articles/parsebio_sketch_integration). Briefly, 5000 representative cells were selected from each pool (based on statistical leverage). Integration was performed on these sketched cells using the RPCAIntegration method. Then, each cell was placed in this integrated space as well using the ProjectIntegration function.
