A Gene Rank Based Approach for Single Cell Similarity Assessment and Clustering

One of the current research directions for single-cell RNA sequencing data is to accurately identify different cell types through unsupervised clustering methods. However, scRNA-seq data analysis is challenging because of their high noise, high dimensionality and sparsity. Moreover, the impact of multiple latent factors on gene expression heterogeneity and on the ability to accurately identify cell types remains unclear. How to overcome these challenges to reveal the true between-cell difference has become the key to the analysis of scRNA-seq data. For these reasons, unsupervised learning for cell populations discovery based on scRNA-seq data analysis has become an important research area. A cell similarity assessment method is the key to accurately identify cell types. Here, we present BioRank, a new cell similarity assessment method that using annotated gene sets and gene rank. In order to evaluate the performances, we cluster cells by two classical clustering algorithms based on the similarity between cells obtained by BioRank. BioRank can be used by any clustering algorithm that requires a similarity matrix. Applying BioRank to twelve published scRNA-seq datasets, the results show that our method is better than or at least as well as several popular similarity assessment methods and single cell clustering methods.