ANU - RSBS - Genomic Interactions - BioInformatics Server

PathExpress

Exploring the metabolic network to interpret post-genomic data

PathExpress is a web-based tool developed to interpret gene expression obtained from post-genomic experiments by identifying the most-relevant sub-networks (meabolic pathways, sub-pathways or enzyme neighbourhood) associated with a subset of genes (e.g. differentially expressed.)

1. Submitting a list of genes

To submit and analyse a list of genes (Fig. 1), it is necessary to:

Select the Genome Array used in the experiment
Paste or upload from a file the List of Identifiers (Affymetrix IDs or Gene Accession Numbers)
Select the type of sub-networks:
- Entire pathways: Comparison of the submitted identifiers to the composition in enzymes of all predefined metabolic pathways from KEGG PATHWAY database.
- Sub-pathways: Comparison of the submitted identifiers to the composition in enzymes of chain of reactions linked to each other by a common compound, within each annotated pathway.
- Enzyme Neighbourhood (EN): Comparison of the submitted list of identifiers to set of enzymes that can be reached in the global metabolic network from an enzyme by traversing a maximum of depth=d compounds.
P-value threshold of significance for hypergeometric distribution test (0.05 by default)
P-value adjustment for multiple hypothesis tests:
- False Discovery Rate (FDR): Determination of the expected proportion of false discovery positive results among all rejected hypothesis
- Bonferroni correction: p-values are multiplied by the number of comparisons (conservative method)

Fig. 1: Submitting a list of genes

2. Assignment

When a list of identifiers has been submitted, PathExpress first assigns them to EC numbers, according to pre-computed relationships. The result page summarizes these assignments (Fig. 2), with:

Genome Array selected
Nb. of identifiers submitted and found in the selected genome array, linked to the original submitted data file
EC assignment: Number of submitted identifiers assigned to EC numbers. The details button displays these assignments (Fig. 3).

Fig. 2: Summary of assignment

2.1. Probe assignment

The detail page of assignment (Fig. 3) displays:

Identifier: Probe set identifier
Public ID: Accession number of a representative sequence
Description: GenBank description associated with the representative public identifier
EC: Enzyme Commission (EC) number, corresponding to the ENZYME entry identifier, that have been assigned to the probe set identifier, linked to the enzyme detail page
Evidence: Cross-link used to assigned the given identifier to the EC number

Fig. 3: Probe assignments

2.2. Enzyme detail

The detail page for a given enzyme (Fig. 4) contains:

Entry: EC number linked to the KEGG database
Name: Recommended name and, if any, alternative names
Reaction(s): List of enzymatic reactions in which the given enzyme is involved
Pathway(s): List of pathways in which the given enzyme is involved
Assignment: List of probe sets assigned to the given enzyme in the given genome array. For each probe set, details are displayed (Identifier, Public ID and description) with the evidence for assigning to the enzyme. Rows are highlighted in yellow if the probe set belongs to the submitted list of genes.

Fig. 4: Enzyme detail

3. Results

PathExpress return as main results the list of sub-networks (complete pathway, sub-pathway or enzyme neighbourhood) associated with the list of identifiers submitted.

3.1. Complete/Metabolic pathways

The main results page shows the list of complete pathways whose enzyme composition intersects with the enzymes corresponding to the list of submitted identifiers, ordered by increasing p-values (Fig. 5). Each row reports information concerning the metabolic pathway and the number of enzymes involved (Nb. of Enzymes). The map button is linked to the representation (graphical and text) of the given pathway (Fig. 6, Fig. 7). The comparison of groups is reported with the number of submitted enzymes involved in the corresponding pathway (Nb. of Enzymes submitted) and the P-value for finding the group per chance, based on the hypergeometric distribution and adjusted according the selected method. The significant pathway, according to the given threshold, are highlighted in red.

Fig. 5: List of metabolic pathways associated with a subset of genes
3.1.1. Graphical representation
Each pathway can be displayed as an automatically graphical representation (Fig. 6). The directed graph contains 2 types of nodes: compounds (labelled with their KEGG identifier and represented as ellipses) and enzymatic reactions (labelled with the EC numbers of the enzymes involved and represented as boxes). Directed arcs between 2 different nodes represent the consumption or the production of compounds by a reaction. Greyed reactions show that the corresponding enzyme has been identifying in the genome array of interest. Reactions mediated bu the enzymes present in the query list are displayed in yellow. The compound name and the reaction formula are displayed as tool-tip when the mouse is over any node. If the user clicks on a reaction node, a new page containing the enzyme description associated with the list of genes assigned is displayed (Fig. 4).

Fig. 6: Graphical representation of a metabolic pathway
3.1.2. List of enzymatic involved
In addition, each patwhay (or sub-network) can be displayed as an enumeration of reactions (Fig. 7). Each rows reports:

Equation: Chemical reaction in the form of an equation
Enzyme(s): Enzyme entries that catalyse the given reaction. Greyed reactions show that the corresponding enzyme has been identified in the genome array whereas enzymes present in the query list are colored in yellow. Each enzyme is linked to its description (Fig. 4).

Fig. 7: List of enzymatic reactions

3.2. Sub-pathways

Using the option Sub-pathways, PathExpress returns the list of chain of enzymatic reactions linked to each other by a common compound within a metabolic pathway, whose composition in enzymes intersect with the list of submitted enzymes. The results page shows the list of sub-pathways, group by pathways (Fig. 8):

Pathway: Name of the pathway the sub-pathway belongs
Entry Point: EC number of the enzyme choose as entry point of the sub-pathway
map: Link to the graphical representation of the sub-pathway, with the entry point highlighted in red
Nb. of Enzymes involved in the sub-pathway
Nb. of enzymes submitted involved in the given sub-pathway
P-value for finding the given EN by chance, based on the hypergeometric distribution, after adjustment (none, FDR or Bonferroni). The significant EN are hightlighter in red, according to the defined threshold.

Fig. 8: List of sub-pathways associated with a subset of genes

3.3. Enzyme Neighbourhood

Using the option Enzyme Neighbourhood, PathExpress returns the list of sub-networks corresponding to sets of reactions in the global network, from a given enzyme (entry point) and for a given depth, order by increasing p-values. Each row reports one EN, compared to the list of submitted enzymes (Fig. 9):

Entry Point: EC number and name of the enzyme used as entry point to explore the enzyme neighbourhood.
map: Link to the graphical representation of the sub-pathway, with the entry point highlighted in red.
Pathway(s) involved: Metabolic pathways represented in the list of reactions that compose the EN.
Nb. of enzymes involved in the EN.
Nb. of submitted enzymes involved in the given EN.
P-value for finding the given EN by chance, based on the hypergeometric distribution, after adjustment (none, FDR or Bonferroni). The significant EN are hightlighter in red, according to the defined threshold.

Fig. 9: List of Enzyme Neighbourhood (EN) associated with a subset of genes