Richard Shipman -- 13AUG2024, Released: 18OCT2024
The Glycoproteomics Graph Tool was developed to create a centralized graph database that captures the intricate relationships in glycoproteomics, linking glycan structures, proteins, genes, and biological processes. By leveraging Neo4j and integrating public data from GlyGen, this tool models glycoproteomics data in a graph format, making it easier to explore and analyze the complex biological interactions of glycoeptides within the context of the proteome, glycome, and beyond.
Graphs are an ideal approach because glycoproteomics sits at the intersection of multiple omics fields (proteomics, genomics, glycomics). This graph tool allows for the integration of multi-omics data, enabling the representation of biological knowledge in a contextualized manner. It facilitates the study of glycoconjugates (glycopeptides) and their connections to key biological entities, enhancing our understanding of glycosylation and its implications in health and disease. The structure makes it highly suitable for future applications in machine learning and artificial intelligence, where understanding relationships between biological components is key.
This GlycoproteomicsGraphTool_release repository serves as the core module in the construction of a glycoproteomics knowledge graph. The primary purpose of this graph is to store glycoproteomic (glycoconjugates) records in relationship to common biological entities and processes. The development of this graph was inspired by the GlyGen workshop titled GlyGen Virtual Training Workshop 2024. (https://wiki.glygen.org/GlyGen_CFDE_Biocuration2024_Workshop) GlyGen data repository also serves as a data source for the glycoproteomics data used in the construction of this knowledge graph.
Graph data model of the GlycoproteomicsGraphTool_release made with the Neo4j provide Arrows.app tool. (https://arrows.app)
The Glycoproteomics Graph Tool leverages Neo4j, a graph database management system, to construct a comprehensive and interactive graph model based on GlyGen data. (https://www.glygen.org/) GlyGen is a data integration and dissemination project for glycomics and glycobiology research, providing extensive datasets on glycans, proteins, genes, and associated biological entities. At the moment, the graph stores glycoproteomics related data from humans and mice.
This tool facilitates the creation and manipulation of a graph database to model complex biological relationships. It uses a series of Cypher queries to load, update, and manage nodes and relationships in the Neo4j graph database. The graph model includes various types of nodes such as proteins, glycans, genes, motifs, and enzymes, as well as relationships that capture the biological interactions between these entities.
The following sections of the readme provide detailed instructions and Cypher queries for setting up constraints, loading data from CSV files, and establishing nodes and relationships within the graph database. This enables researchers to explore and analyze the intricate web of interactions in glycomics data, thereby advancing our understanding of glycan structures and their biological roles.
This graph acts as a core for glycoproteomics projects, with GlyGen as the primary data source. In the future, additional data sources may be integrated to expand the coverage of proteins, glycans, and beyond to other related compounds and reactions. The stored graph topology of glycoproteomics data could also support applications in ML/AI. Additional glycoproteomes from other species could be added as well. Note: Curation steps are also needed find and fix errors in the graph, current release does contain bugs that will be resolved in future releases.
-
Download the GlycoproteomicsGraphTool_release github repository with required ZIP File containing CSVs and neo4j.config.
-
Extract and Place CSV Files in the Import Folder of your Neo4j Database
Extract the contents of the ZIP file (graph_data.zip). Move all the CSV files into the import directory of your Neo4j database. The import directory is typically located within the Neo4j installation directory. This step is crucial for Neo4j to access the data files directly.
- Copy the neo4j.config File From the data Folder Into the Neo4j Config Folder
Or add the following parameters to your neo4j.config file with a text editor at the bottom of the file:
dbms.memory.heap.initial_size=2G
dbms.memory.heap.max_size=8G
dbms.memory.pagecache.size=4G
Note: System requirements: 8Gb+ RAM.
- Run Cypher Queries in the core_graph_statements.cypher File in the Neo4j Browser
Open the Neo4j browser from the App or the web.
Follow the order of the Cypher queries as outlined in the Cypher file (core_graph_statements.cypher). Each section corresponds to a specific type of node or relationship within the graph model. Execute these queries in your Neo4j browser or using a Neo4j client.
- Explore the Glycoproteomics Graph Tool
Use the following use cases and test examples to explore the knowledge graph.
- Display groups of glycans that are related to one another by composition and structure and display monosaccharide and has linkage subgraph.
MATCH p=(g1:Glycan)-[r:has]-(g2:Glycan)-[r1:has_structure]-()-[r2:linkage*]-()
RETURN p LIMIT 100
- Test the following cypher to see if glycan to monosaccharide relationships loaded properly.
MATCH p=(g:Glycan {glytoucan_ac: 'G00025HU'})-[r:has_structure]-()-[r2:linkage*]-()
RETURN p;
- Which enzymes synthesize this set of glycans?
UNWIND ["G00025YC", "G00026MO", "G00030MO"] AS value
MATCH path=(g:Glycan{glytoucan_ac: value})-[]-(e:Enzyme)-[]-(p:Protein)-[]-(ge:Gene)
return path limit 100
- Greater than 2 dHex in composition. dHex =< 2.
UNWIND ["2", "3", "4", "5"] AS value
MATCH path=(g:Glycan{dHex: value})-[]-(e:Enzyme)-[]-(p:Protein)-[]-(ge:Gene)
return path limit 100;
- find all glycans from humans that are synthesized by FUT8
MATCH p=(n:Enzyme{species:'Homo sapiens', gene_name:'FUT8'})-[]-(g:Glycan{glytoucan_type:'Saccharide'})
RETURN g.glytoucan_ac;
g.glytoucan_ac
"G32212NQ"
"G00888JR"
"G88365LW"
"G24782UB"
"G62145FH"
...
- Map Glycogenes and Enzymes to Chromosome ID (Gene to Chromosome Mapping) “Which genes of Homo sapiens are associated with proteins that are also enzymes, and what are their gene names, chromosome IDs, and species, ordered by chromosome ID in ascending order? Which chromosomes are glycogenes of interest found on?”
MATCH x=(g:Gene{species:'Homo sapiens'})--(p:Protein)--(e:Enzyme)
RETURN g.ensembl_gene_id,
p.uniprotkb_canonical_ac,
p.gene_name,
g.chromosome_id,
g.start_pos,
g.species
ORDER BY g.chromosome_id ASC,
g.start_pos;
g.ensembl_gene_id p.uniprotkb_canonical_ac p.gene_name g.chromosome_id g.start_pos g.species
"ENSG00000158850" "O60512-1" "B4GALT3" "1" "161171310" "Homo sapiens"
"ENSG00000162630" "O43825-1" "B3GALT2" "1" "193178730" "Homo sapiens"
"ENSG00000143641" "Q10471-1" "GALNT2" "1" "230057990" "Homo sapiens"
"ENSG00000184389" "U3KPV4-1" "A3GALT2" "1" "33306766" "Homo sapiens"
"ENSG00000126091" "Q11203-1" "ST3GAL3" "1" "43705824" "Homo sapiens"
- Which glycans are associated with protein uniprot accession P00450-1 (Ceruloplasmin)?
MATCH x=(p:Protein{uniprotkb_canonical_ac:'P00450-1'})--(gs:GlycosylationSite)--(gc:Glycoconjugate)--(g:Glycan)
RETURN g.glytoucan_ac 
g.glytoucan_ac
"G00912UN"
"G94917XT"
"G88374WZ"
"G10486CT"
"G15038BD" ...
GlyGen: Computational and Informatics Resources for Glycoscience, Glycobiology, Volume 30, Issue 2, February 2020, Pages 72–73, https://doi.org/10.1093/glycob/cwz080
This project is licensed under the GPL-3.0 license.