Subpopulation Identification from Single-Cell RNA-seq Data

To develop algorithms to jointly consider imputation and clustering analysis for single-cell RNA-sequencing data which are zero-inflated.

During this project, the Qiu lab developed a novel co-occurrence clustering algorithm that does not depend on any preprocessing, such as normalization, imputation or dimension reduction Specifically, this algorithm differs from existing methods as it embraces dropouts as a useful signal for defining cell types. This work is further explored in their openly available preprint.