Development and Applications of Ultrasensitive Methods for Single-Cell Analyses

Single-cell proteomics is a rapidly advancing and increasingly vital field that offers the potential to uncover phenotypically direct aspects of cellular heterogeneity through the comprehensive analysis of proteins at the single-cell level. This technique provides unprecedented insights into cellular heterogeneity, revealing the diverse functional states of individual cells within complex tissues, especially when combined with other omics in a multi-omics analysis. While advancements in mass spectrometry, sample processing, and data analysis have improved the state of single-cell analyses, significant challenges remain, derived from the limited amount of analytes in a single cell and subsequent sample loss from container binding. This thesis provides a comprehensive overview of our efforts to address some of the significant challenges of single-cell analyses by reducing the amount of sample loss, improving the throughput, and improving the detectability of peptides. To address the significant issue of sample loss, we have created a device to process the single-cell samples that we termed the “Levcell.” The Levcell uses acoustic levitation to levitate up to 6 samples in the air without needing a container, eliminating the sample loss from container binding. Combined with a custom-built single-cell isolation system, reagent sample addition arm, and computer vision, the Levcell is a comprehensive tool to process single-cell samples in an automated and high-throughput fashion. Additional uses for the Levcell are also explored in this thesis. Additionally, we have developed a high-throughput dual-column system to improve LCMS utilization time efficiently. This system alternates between two columns to wash, equilibrate, and load a subsequent sample in one column, while the current sample is being assessed in the other, significantly reducing the time the LC-MS does not detect the sample. When applied to single-cell analyses, the dual-column system can efficiently process the samples needed to identify cellular heterogeneity. Lastly, we comprehensively analyze improving peptide detectability through proteomewide alkylation. By labeling peptide samples from cell digests with aldehydes of various side chain lengths, we systematically test for biases in labeling sites, improvements in peptide intensity, and finally, labeling efficiencies of reactions. Our findings could lead to the applications of improving identifications of low-yield samples down to that of a single cell.