We describe a methodology for partitioning scRNA-seq datasets into metacells: disjoint and homogenous groups of profiles that could have been resampled from the same cell. Unlike clustering analysis, our algorithm specializes at obtaining granular as opposed to maximal groups. |