Skip to content

biohackathon-japan/bh23-multiomics-knowledge-graph

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

37 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Multi-omics data integration on knowledge graphs

Objective

Integrate multiomics data using knowledge graph and Graph Convolutional Networks

Team

  • Alberto Labarga
  • María Gomez
  • Robert Hoehndorf
  • Nuria Queralt

Outcomes

Biomedical knowledge graph

We have integrated the following data resources

  • Uniprot
  • mirDB
  • MED-RT
  • DisGeNet
  • Rhea
  • Drugbank (WIP)

to generate a biomedical knowledge graph

image

An updated visual description of the Knowledge Graph is here

The workflow and queries for constructing the knowledge graph is available here.

Omics data

  • Biostudies mapping to knowledge graph
  • Multiomics data RDF representation definition

Clinical knowledge graph

Machine Learning on knowledge graph

For our use case, we integrated the different omics and clinical data with the knowledge graph to create a networkx-based graph representation. We then explore different Graph Neural Networks libraries that could be used for the different biological questions we had. The libraries we used where

- StellarGraph: https://github.com/stellargraph/stellargraph#installation
- KGCN: https://github.com/clinfo/kGCN/

Other technologies leveraged were:

  • SPARQLWrapper
  • RDFlib
  • Networkx

The algorithms proposed for the different use cases are:

- Patient classification: Graph classification
- Biomarker prediction: Node classification
- Data imputation: Node feature prediction
- Drup reporpousing: Link prediction

and the notebooks are available at the src/gcn/notebooks folder.

Check https://github.com/alabarga/biohackathon-jp-2023/tree/main/multiomics-knowledge-graph for a more detailed view of work done during Biohackathon

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages