Mathematical Modeling of N-Glycosylation: Model Construction and Parameter Estimation from Mass Spectra of Chinese Hamster Ovary Cell Mutants
MetadataShow full item record
Colossal advances in molecular cellular biology have witnessed the growth of genomics, proteomics and other "–omics" fields, built on reductionist approaches aimed at annotating a vast array of biological components, from genes and proteins to entire cells and tissues. In recent years, these advancements have propelled the field of glycomics—designed to explore the role of carbohydrates in a wide variety of biological processes—forward at a rapid pace. However, the multi-substrate nature of glycosylation enzymes leads to a large number of observed heterogeneous oligosaccharide structures present on proteins. The distributions of these structures are difficult to control via traditional knockout/over-expression approaches. In order to better identify which enzymes regulate key oligosaccharides, we analyzed mass spectra data from Chinese Hamster Ovary cells and their respective mutants. Subsequently, we constructed in silico reaction networks that generate these oligosaccharide structures and attempted to fit the distribution obtained from analysis of the mass spectra. This fit was determined by minimizing a `fitness function' which describes the difference between experimental data and the simulation results. In the computer simulations, the introduction of the appropriate biochemical change only qualitatively reflected the changes observed in glycan fractions. Despite this, the matching of wildtype and mutant glycome profiles were not satisfactory. Improvements and strategies to obtain better data fits are discussed here as well as the future implications of constructing the current model.