Single-Virus/Single-Cell Sequencing with Droplet Microfluidics

Abstract

Droplet microfluidics offers numerous advantages for biological applications. One of the most profound characteristics is the ability to encapsulate materials into picoliter droplets in a high throughput manner. Since high-resolution analyses of cells or molecules with conventional wet lab techniques are labor-intensive and time-consuming, the throughput and automation offered by droplet microfluidics make it an ideal tool for biological applications involving large numbers of biological entities. In contrast, microfluidic devices can process thousands to millions of molecules or cells within ten minutes and with minimal manual control. In this dissertation, I present the method development of three biological applications, all of which benefit greatly from the high-throughput processing of droplet microfluidics. In addition, each of them takes advantage of some other aspects of droplet microfluidics, presenting robust methods for solving questions in viral discovery, single-cell co-detection of RNA and protein, and targeted protein evolution. I first describe the development of a high-sensitivity method for recovering the complete genome sequence of a single virus from an environmental sample. Characterizing unknown viruses is vital for understanding and preventing infectious diseases; however, existing methods with cultured-based models or bulk-sequencing of environmental samples fail to effectively recover their genome sequences for analysis. We incorporate an RNA-based gene detection method with whole genome amplification in droplets to enrich the genome of any specific viral species in a heterogeneous sample. Additionally, we achieve an ultra-high sensitivity for genome recovery, offering a solution to the challenging problem of recovering rare viruses. In the next chapter, I present a method for the simultaneous quantification of cellular RNA and an intracellular protein in single cells and apply it to study BK virus infection. BK virus is a significant risk factor for immunosuppressed individuals; understanding the cellular responses it induces is critical for the development of targeted therapies. VP1 is an intracellular viral protein indicating the degree of infection of each cell. Measuring the level of VP1 along with the cellular transcriptome is needed to analyze molecular pathways controlling the infection. We develop a high-throughput method to barcode cellular RNA and VP1 protein in single cells and quantify their expressions with sequencing analysis. In the work described in Chapter 4, droplet microfluidics is applied to directed protein evolution, where large numbers of variant sequence-activity paired are generated for building machine-learning models. We encapsulate cells with gene variants in droplets, measure their activity with a fluorescent-based assay, and recover the variant sequences with sequencing. Various statistical models are tested for optimal prediction of variant activity from amino acid sequences.