A downsampling method enables robust clustering and integration of single-cell transcriptome data

•We propose a model-based downsampling algorithm for single-cell RNA sequencing analysis, the minimal unbiased representative points (MURPXMBD).•MURP defines a set of representative points by reducing gene-wise random, independent errors, while best retaining the covariance structure of biological origin.•MURP provides unbiased representation of the original cell population, which is robust to highly imbalanced representation of cell types and batch effects, thus improve the clustering proficiency and downstream integration analysis.