Data-Driven Mass Spectrometry for Drug Discovery and Bioanalysis: High-Throughput Desi and 2D MS/MS

Mass spectrometry (MS) stands out among modern analytical techniques due to its unique combination of rapid, precise measurement capabilities and broad applicability across scientific disciplines. From its origins in physics to its widespread adoption in chemistry and biology, the development of MS can be understood through the dual lenses of data and information: on one hand, advances in instrumentation and ionization methods have expanded the scope and quality of data that can be acquired; on the other, innovative experimental designs and computational approaches have enhanced the ability to extract meaningful chemical information from that data. This thesis investigates the current frontier of MS development along both dimensions, with a particular focus on desorption electrospray ionization (DESI) and two-dimensional tandem mass spectrometry (2D MS/MS). 

The development of DESI as an ambient ionization method has significantly expanded the data-generating capacity of MS by enabling direct sample analysis with minimal preparation or separation. In this work, we described a series of experiments that leveraged the high-throughput capabilities of the DARPA Purdue Make-It platform, particularly its exceptional MS acquisition rate approaching 1 Hz. Details on experimental designs and strategies for managing, analyzing, and extracting meaningful information from the resulting large-scale datasets are described for two major applications: (1) the creation of large-scale MS/MS libraries to explore fragmentation characteristics specific to drug-like small molecules, and (2) the acceleration of screening workflows for synthetic reactions, enzymatic transformations, and bacterial cultures in support of drug discovery and development – one of the most impactful applications of high-throughput experimentation in applied chemistry. 

A key strength of MS lies in its MS/MS capabilities, which enable structural characterization of a wide range of chemical and biochemical molecules. However, as MS applications expanded to more complex analytes, traditional MS/MS acquisition methods proved inefficient in terms of information yield. Using 2D MS/MS, a novel technique capable of rapidly reconstructing precursor-product relationships of all ions within a mixture without precursor isolation, we developed a method for comprehensive structural characterization of biopolymers. By combining in-source fragmentation with 2D MS/MS, this method maximizes the depth of structural information obtainable in a single experiment. Specifically, MSⁿ fragmentation pathways can be reconstructed, providing structural insights previously inaccessible through conventional approaches. To further extend the method’s applicability to modern omics workflows, we developed a graph theory–based data analysis framework that enables efficient data organization, automated deconvolution, and de novo peptide sequencing, without requiring extensive chromatographic separation or large spectral databases. 

Collectively, this work seeks to contribute to the ongoing progress of mass spectrometry toward faster, broader data acquisition and deeper, more insightful information extraction, with the aim of supporting MS’s continued expansion across chemical disciplines.