Different cell-types is one of the most fundamental aspects of multi-cellular organisms. Traditionally, cell-types are defined by morphological properties or surface markers. Single-cell RNA-seq experiments have made a more rigorous approach possible. Given a sample of cells from a tissue, one can use unsupervised clustering to identify groups of cells, i.e. cell-types, with similar expression profiles.
Due to the high levels of noise and the high-dimensionality of the transcriptome, clustering cells remains a challenging task. SC3 is a user-friendly computational tool which allows the user to explore different clustering options. Furthermore, SC3 identifies genes which are specific to each cluster as well as possible outliers. Importantly, SC3 includes a semi-supervised mode which means can be used for large-scale drop-seq experiments.