GlycoDiveR: a modular R framework to analyze and visualize highly dimensional glycoproteomics data

Abstract

Mass spectrometry-based glycoproteomics is a critical platform for understanding the complex roles of protein glycosylation in biological systems, yet visualizing multidimensional glycoproteomics datasets remains a significant bottleneck in data interpretation and communication. Glycan microheterogeneity, i.e., the potential for a glycosite to be modified by multiple glycans, defies the binary presence-absence logic used in analyses of other post-translational modifications. Instead, glycoproteomics necessitates intentionally designed data structures and visualizations that are glycoform-centric, not just site-centric. Additionally, there is a need for complementary degrees of data analysis that alternate between glycoproteome-scale patterns and glycosite-specific regulation. Several bespoke frameworks for visualizing glycoproteomics data have emerged, but they often require advanced programming expertise and are designed for a single study rather than broad application. Here, we present our efforts to harmonize post-search data analysis of glycoproteomics through a modular R framework called GlycoDiveR. This platform streamlines import, transformation, and curation of qualitative and quantitative glycopeptide identifications, including support for raw output from multiple search engines. GlycoDiveR is designed to integrate seamlessly into existing analysis workflows by enabling fast, flexible exploration of highly dimensional glycoproteomics datasets via a consistently formatted data architecture. Our goal is to offer a customizable set of glycosylation-specific visualizations with minimal coding, while keeping data accessible to users who wish to further customize their characterization strategies. It also maintains a modular design that supports the continual addition of visualizations, analyses, and export functions. Ultimately, GlycoDiveR is meant to improve accessibility of glycoproteomic-specific analyses and lower the barrier to exploring biological narratives embedded in rich glycoproteomic datasets. GlycoDiveR is open-source and freely available at https://github.com/riley-research/GlycoDiveR.