Lecture 15&16: Graph Theory and Systems Biology Flashcards
How are Gene regulation networks expressed?
Nodes: Represent transcription factors (TFs) and controller genes. Transcription factors are proteins that regulate the expression of other genes
Edges: Represent gene regulatory interactions. These interactions can be either activation or repression.
Gene regulation is hierarchical - where a few master regulator genes control multiple lower level genes.
How are metabolic networks regulated?
- Signaling network with proteins (de-) phosphorylate other proteins to transmit a signal.
- Gene regulation network with transcription factors bind to particular sequences in the genomic DNA to turn on/off other genes.
- This then increases/decreases the expression of particular enzymes which changes the processing of small molecules in metabolic pathways
What is the KEGG database?
A comprehensive database that integrates information about biological pathways, genes, diseases, and drugs.
It is widely used in bioinformatics and systems biology for studying interactions within cellular systems.
How do proteins interact with each other?
(De-)phosphorylation or other
modifications of amino acid
residues (in signaling pathways)
Allosteric activation (transcription
factors, enzymes)
Inhibitory binding (covering the
active site)
Physical closeness of enzymes
along metabolic pathway.
What are the properties of graphs when it comes to graph theory?
Node: A vertex which represents an entity that we wish to model that can have a defined relationship with other nodes
Edge: A connection between two vertices (nodes) that specifies some relationship between the vertices.
Adjacency: Two nodes are adjacent if connected by an edge
What are some of the graph topologies?
Directed: Edges have directionality
Undirected: Edges have no directionality
Cyclic: Possible to follow edges back to nodes
Acyclic: Not possible to return to nodes
Connected: All nodes connected together
Tree: There is only 1 path between any pair of nodes
Rooted tree: One root, where all edges are spouted ut from.
What are some common algorithms in graph theory?
Graph Traversal: Systematic exploration of a graph by visiting nodes and edges.
Measure distance: Calculates the shortest path between two nodes.
Centrality: Determines the most “important” nodes in a network.
Bi-connected sub-graphs: Finds sub-networks where removing any single node does not disconnect the network.
Clique/sub-graph matching: Identifies complete sub-networks (cliques) where every node is connected.
Bridge detection: Identifies edges whose removal would disconnect the graph.
Cycle detection: Identifies circular paths where a node can be revisited.
What are some common algorithms and their biological implication?
Graph Traversal: Enumerate all synthesis paths for a metabolite in a network
Measure distance: How closely related are two organisms
Centrality: Detect critical signaling proteins
Bi-connected sub-graphs: Find fault resilient sub graphs in signaling networks.
Clique/sub-graph matching: Fast detection of protein structures
Bridge detection: Find fault prone regions in signaling networks
Cycle detection: Find feedback loops in expression networks
What are the steps in a context likelihood of relatedness algorithm?
- Gather all transcription data
- Calculate mutual information between expression levels of all pairs of genes
- Build an MI matrix
- Calculate a z-score for each putative TF and putative target.
- Calculate a joint z-score
- Any joint z-score over a particular threshold should be accepted as regulation.
