Expanding the scope and utility of single-cell genomic technologies

Abstract

Technological advancements in single-cell genomic technologies have led to an exponential increase in number of cells from which biologists are able to measure various molecular profiles. This dramatic increase in scalability opens up the potential to leverage such technologies as the basis for a wide array of off-label applications. For example, the genetic screening field has typically been limited approaches that examine changes in relative abundance of genotypes represented in a given cell population before and after screening/selection as a proxy for phenotype. Molecular measurements such as RNA sequencing (RNA-seq) could allow a more generic and direct readout of molecular phenotypes, but performing bulk RNA-seq on many mutants in an arrayed format dramatically limits scalability. Single-cell RNA-seq, if modified to enable the readout of cell genotypes, would allow us to examine the molecular changes resulting from individual genotypes in a pooled format. In fact, single-cell technologies create the tremendous opportunity of multi-modal measurements (e.g. gene expression in addition to genotypes, lineage information, protein abundances via oligo tagged antibodies, T and B cell receptor sequences, entire additional molecular measurements like chromatin accessibility). An important step in the progression towards this goal is the adaptation of other molecular profiling techniques beyond RNA sequencing to single-cell formats, which come with their own experimental and analytical challenges that have yet to be solved within the field. In this dissertation, I first introduce an attempt to use single-cell transcriptomics to serve as a readout for genetic screens. In Chapter 2, I introduce our work to enable such a readout from CRISPR and CRISPRi-based genetic screens and find that a number of approaches that have been used to tackle this problem suffer from high error rates (~50\%) in the assignment of correct genotypes to cells. Ultimately we recommend a particular existing design that does not suffer from this common flaw. In Chapter 3, we utilize this method in conjunction with highly multiplexed perturbations to enable the screening of almost 6000 putative regulatory elements in the genome, measuring their impact on gene expression in cis. Notably an experiment of this scale would have been extremely difficult to achieve without the use of single-cell technologies. Lastly, in Chapter 4, I shift focus to our efforts to expand the set of molecular measurements that we can make with single-cell technologies. Specifically, I describe our efforts to measure chromatin accessibility using single-cell ATAC-seq across 13 different tissues in 8-week old mice and how we went about addressing the many computational challenges that arise when attempting to analyze and interpret such datasets.