An automated software-assisted approach for exploring metabolic susceptibility and degradation products in macromolecules using High-Resolution Mass Spectrometry

Abstract

A comprehensive understanding of drug metabolism is crucial for advancements in drug development. Automation has improved various stages of this process, from compound procurement to data analysis, supporting small molecules, peptides, and oligonucleotides. However, challenges remain, particularly in the time-consuming analysis of samples for metabolite identification. This article introduces new algorithms for automated Liquid Chromatography-High-Resolution Mass Spectrometry (LC-HRMS) data applicable to both small and macromolecules. While methodologies for small molecules are well established, adapting them for macromolecules presents challenges, including computational demands, peak detection complexities, and visualization issues. A data analysis employing diverse algorithms in the data preprocessing step was conducted across six datasets, ranging from small/medium linear or macrocyclic peptides to oligonucleotides with natural and unnatural monomers. Two peak detection approaches were evaluated: using the monoisotopic mass versus the most abundant isotope for mass calculation. Additionally, an exploration of two distinct structure visualization options was conducted for one of the datasets. Furthermore, data obtained through two different acquisition modes was processed. The computational time required for data processing was recorded throughout, ranging from 5 minutes to 2 hours per experiment. The results have been compared against prior studies, revealing substantial reductions in processing time, consistent identification of degradation products, and improved visualization techniques, thereby enhancing result interpretation. A comprehensive identification of 970 metabolites was achieved under varied incubation conditions across the six datasets, showcasing the workflows efficiency in managing experimental data within a molecular range from 700 to 7630 Daltons (Da). Particularly in larger molecules, the most abundant mass algorithm demonstrated higher scores and a greater number of matches, instilling greater confidence in the accurate prediction of metabolite structures. It has been illustrated how the visualization algorithm for macromolecules allows the combination of monomer and atom/bond notation, facilitating a clear depiction of metabolic changes in the molecular structure.