Bioinformatics
10. Biological networks
Content: networks, cytoscape
10.1 Networks
- represent relations between proteins, genes, transcription factors,
diseases, drugs, …
- relation can be represented in a table OR better visualized with a network
Networks as Tools
- analysis: topological properties, hubs and subnetworks, classify, cluster
and diffuse, data integration
- visualization: data overlays, layouts and animation, exploratory analysis,
context and interpretation
Applications in research
Some basic graph theory
- graph: mathematical way of representing network
o can be directed or undirected
directed graphs: cyclic or acyclic
o can be weighted or unweighted
analytical approaches
- levels of organization of complex networks
o node degree (= how many edges) provides information about
single nodes
o ≥ 3 nodes represent a motif
o >> noedes: modules or communities
- Hub: node with very high degree
Biological network taxonomy
- Interactions: prot-prot, prot-ligand, domain-domain, others (residue or
atomic, cell-cell, epidemiology, social networks)
- Pathways: metabolic, signaling, regulatory, disease
- Key distinctions: pruned and curated, mechanistic and contextual details
- Similarity: protein/sequence similarity network (PSNs or SSNs), chemical
similarity, ligand similarity (SEA), others (tag clouds, topic maps)
Protein-protein interaction data
- Data sources
o Experimental techniques
High throughput
Low throughput
o Computational techniques
o Public repositories
, Experimental techniques
- high throughput
- low throughput
computational techniques
- text mining: search literature for references to protein-protein interactions
- orthology: predicts interaction based orthologous pairs in another species
- domain pairs: predicts interaction based on domains interaction in other
proteins
10.2 Cytoscape
- cytoscape can import network data from
o files (or URLs)
o public repositories
o automation
NDEx (CyNDEx App)
Data mapping
- mapping of data values associated with graph elements onto graph visuals
- mapping types: continuous (numeric values), discrete (categories),
passthrough (labels)
saving and exporting
11. Gene Ontology
11.1 Gene ontology
- ontology: formale representation of a set of concepts within a domein and
the relationships between those concepts
- structured & controlled vocabulary
- goals: limit complexity & organizing information
- originally constructed for a very specific purpose: annotation of genes and
proteins in genomic and protein databases
- many other applications arose afterwards
o gene set enrichment analysis
o developing automated ways of deriving information about gene
function
- development is annotation driven
o terms added as a new species annotated & as-needed basis
GO structure
- GO terms divided into three parts
o Cellular component (CC)
o Molecular function (MF)
A single reaction or activity
10. Biological networks
Content: networks, cytoscape
10.1 Networks
- represent relations between proteins, genes, transcription factors,
diseases, drugs, …
- relation can be represented in a table OR better visualized with a network
Networks as Tools
- analysis: topological properties, hubs and subnetworks, classify, cluster
and diffuse, data integration
- visualization: data overlays, layouts and animation, exploratory analysis,
context and interpretation
Applications in research
Some basic graph theory
- graph: mathematical way of representing network
o can be directed or undirected
directed graphs: cyclic or acyclic
o can be weighted or unweighted
analytical approaches
- levels of organization of complex networks
o node degree (= how many edges) provides information about
single nodes
o ≥ 3 nodes represent a motif
o >> noedes: modules or communities
- Hub: node with very high degree
Biological network taxonomy
- Interactions: prot-prot, prot-ligand, domain-domain, others (residue or
atomic, cell-cell, epidemiology, social networks)
- Pathways: metabolic, signaling, regulatory, disease
- Key distinctions: pruned and curated, mechanistic and contextual details
- Similarity: protein/sequence similarity network (PSNs or SSNs), chemical
similarity, ligand similarity (SEA), others (tag clouds, topic maps)
Protein-protein interaction data
- Data sources
o Experimental techniques
High throughput
Low throughput
o Computational techniques
o Public repositories
, Experimental techniques
- high throughput
- low throughput
computational techniques
- text mining: search literature for references to protein-protein interactions
- orthology: predicts interaction based orthologous pairs in another species
- domain pairs: predicts interaction based on domains interaction in other
proteins
10.2 Cytoscape
- cytoscape can import network data from
o files (or URLs)
o public repositories
o automation
NDEx (CyNDEx App)
Data mapping
- mapping of data values associated with graph elements onto graph visuals
- mapping types: continuous (numeric values), discrete (categories),
passthrough (labels)
saving and exporting
11. Gene Ontology
11.1 Gene ontology
- ontology: formale representation of a set of concepts within a domein and
the relationships between those concepts
- structured & controlled vocabulary
- goals: limit complexity & organizing information
- originally constructed for a very specific purpose: annotation of genes and
proteins in genomic and protein databases
- many other applications arose afterwards
o gene set enrichment analysis
o developing automated ways of deriving information about gene
function
- development is annotation driven
o terms added as a new species annotated & as-needed basis
GO structure
- GO terms divided into three parts
o Cellular component (CC)
o Molecular function (MF)
A single reaction or activity
