Aligning heterogenous single cell assay datasets

Abstract

Pluripotent stem cells offer strong promise for regenerative medicine but the pluripotent cell state is poorly understood. The goal of this thesis is the development of methods to analyze how the multiple facets of cell state-including gene expression, chromosome contacts, and chromatin accessibility-relate in the context of stem cells. The variability of each of these characteristics cannot be deduced from population studies, and while recent advances in single-cell transcriptomics have led to the development of a number of different single-cell assays, datasets that collect multiple types of assays on the same cells are rare. In this thesis, we explore the ability of three methods to integrate datasets from different single-cell assays based on an existing paired single-cell dataset of ATAC-seq and RNA-seq for human A549 cells. We then apply these methods to map the variability between three single-cell datasets-ATAC-seq, RNA-seq, and Hi-C-on pluripotent mouse embryonic stem cells and assess the performance of these methods.