Single-Cell Gene Set Analysis with Applications in Tumor Heterogeneity

• Activity Number: 178 - Statistical Methods for Analysis of Heterogeneous Tissue Samples in Bulk and Single-Cell Sequencing Data
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Lingling An* and Di Ran and Nicholas Lytal
• Companies: University of Arizona and University of Arizona and University of Arizona
• Keywords: single cell sequencing; gene set analysis; pathway

Abstract

The advance of single-cell RNA-sequencing provides the capability to detect transcriptional heterogeneity within a tumor tissue at single cell resolution. However, characterization of heterogeneity by using this new technology remains challenging due to a high level of noise. Here we present new statistical methods to identify cell subpopulations by finding significant pathways/gene sets that are associated with cell-to-cell variabilities. Each of such gene sets could provide a certain amount of evidence distinguishing cell-to-cell differences and group cells into subpopulations. After filtering unwanted gene sets, ensemble clustering method is applied to integrate each individual clustering result and give a final ensemble matrix, which summarizes the possibility that each pair of cells is in the same subgroups. By using single-cell RNA-sequencing data with known cell labels, we demonstrate that our method can characterize correct subtypes in primary glioblastoma and reveal pathways with sound biological insights. The results also show the potential possibility of prognostic and therapeutic implications of intratumoral heterogeneity.