(188dm) Integrative Analysis of Glycosylation Networks Using Transcriptomics and Glycomics Data Sets | AIChE

(188dm) Integrative Analysis of Glycosylation Networks Using Transcriptomics and Glycomics Data Sets


Zhou, Y. - Presenter, SUNY at Buffalo
Neelamegham, S., University at Buffalo, State University of New York
Groth, T., State University of New York-Buffalo

Integrative analysis of
glycosylation networks using transcriptomics and glycomics data sets


, Theodore Groth and Sriram Neelamegham

Chemical and Biological
Engineering, University at Buffalo, SUNY

Email: yusenzho@buffalo.edu, neel@buffalo.edu

glycome constitutes the entire complement of glycans and glycoconjugates that
compose a cell, tissue or organism. With the advent of new high-throughput technologies,
vast amounts of data related to the glycome are being collected at the
transcript level using RNA-seq, and structure level using mass spectrometry
(MS). These data are currently generated by multiple laboratories and collated
in various online databases. The relation between data collected using
different modalities is unknown since there is no streamlined computational
method to integrate knowledge from these different experimental techniques. To
address this limitation, we developed an online tool called GNAT-Web
(Glycosylation Network Analysis Toolbox-Web). This resource contains a custom,
curated database for mammalian glycosyltransferases and glycosidases based on
computer-generated queries to the Uniprot, Brenda, KEGG, and Japanese GlycoGene
repositories. These online data are complemented by manual-expert annotation of
glycosylation pathway reaction rate laws using primary biochemical literature.
The annotated pathways analyze all aspects of cellular glycosylation:
N-/O-linked glycan biosynthesis, glycosphingolipid construction, glycosaminoglycan
assembly and related metabolic pathways. Using knowledge of the existing glycoenzymes
in a system based on RNA-Seq experiments, GNAT-Web performs ‘forward inference’
to automatically predict/constructs all possible glycans that may be created in
a given system. Using nano-LC-MS/MS glycomics profiling data and a ‘reverse
inference’ algorithm, GNAT-web also assembles putative pathways existing in
cells based on structural information. Kinetic network modeling is then used to
reconcile the differences between data collected using different experimental
modalities, in order to construct the ‘optimized cellular glycosylation
network’. Experimental analysis using a panel of human leukocyte transcriptomics
data and corresponding MS results are used to demonstrate the functionality and
utility of GNAT-web. Such analysis shows that human leukemic cells contain
pathways that result in the formation of extended lactosamine extensions
bearing sialylofucosylated glycans, along with O-GalNAc, O-fucose and O-mannose
type O-glycosylation pathways, and Lacto series glycosphingolipid biosynthetic
pathways. Overall, GNAT-Web is a useful tool to integrate systems level
knowledge generated using different experimental modalities in order to simplify
the presentation of complex cellular glycosylation pathways.