I've been reading literature on differential expression of scRNA-Seq data, and most of the comparison studies refer to using count data as many of the DE methods (for bulk and single cell) require integer values. However, when obtaining the "count" matrix from GSE (for example), not all of the values are integers. Is there some generalized/accepted method (such as rounding or using the ceiling function) to make these non-integer values contained in the "count" matrix into integers so that they can be input into these DE methods?