engleski

Modeling IgG glycosylation changes in aging

Most membrane and extracellular proteins are glycosylated, which makes the attached oligosaccharide an integral part of the glycoprotein, thus altering its structure and function. Glycosylation of IgG antibodies is of particular interest. The glycan part critically affects the effector function of the antibody. The glycosylation pattern is a result of a complex interplay of many enzymes connected into an intricate network. In order to understand how this network changes with age, we used the IgG glycome data from several hundreds of individuals. We built a correlation network of glycan structures based on our extensive dataset. Our model uses covariance regression with Markov chain Monte Carlo simulations, with independent fitting of the mean and the covariance for each glycan structure. This strategy required careful model selection, evidence of Markov chain convergence and visualization of the covariance matrix in a network form, which represents the basis for biological interpretation. We found good clustering into different IgG classes when the model was applied to the integral dataset, with strong positive correlations within individual IgG classes and mostly negative correlations between different classes. To gain a deeper insight into the aging process, we shifted our focus to the correlations within individual IgG classes. Our preliminary results have identified several key glycosylation enzymes that might serve as control points for the whole network and point to their role in changes in the glycome observed during aging, affecting the IgG effector function and thus having a possibly important role in both the autoimmune diseases and cancer.

glycosylation; IgG; aging; MCMC; network

nije evidentirano