Pathway Analysis

1
Pathway Analysis
Martina Kutmon
2
Contents

• Background on Pathway Analysis
• Data Analysis with PathVisio
• Introduction to the Afternoon Session

3
Biological Pathways
4
Why Pathway Analysis?

• Intuitive to biologists
• Puts data in biological context
• More intuitive way of looking at your data
• More efficient than looking up gene-by-gene
• Computational analysis
• Overrepresentation analysis
• Network analysis

5
Why Pathway Analysis?
6
Biological Context

• Statistical results
• 1,300 genes are significantly regulated after
  treatment with X
• Biological Meaning
• Is a certain biological pathway activated or
  deactivated?
• Which genes in these pathway are significantly
  changed?

7
Pathway Collection

• Where to get pathways?
• Online pathway databases
• WikiPathways www.wikipathways.org
• Reactome www.reactome.org
• Many more ... http//pathguide.org

8
Identifier Mapping
Identifier Mapping
Annotation ENSG00000131828
9
Identifier Mapping

• Microarrays typically use internal ids
• Affymetrix 205749_at
• Agilent A_14_P106416
• Illumina ILMN_4380
• Pathways typically use gene/protein ids
• Entrez Gene 1543
• Ensembl ENSG00000140465
• UniProt P04637

10
Identifier Mapping

• 2 scenarios
• Software will take care of it
• e.g. PathVisio uses synonym databases
• You will have to convert the ids yourself
• DAVID http//david.abcc.ncifcrf.gov
• SOURCE http//smd.stanford.edu/cgi-bin/source/so
  urceBatchSearch
• BioMART http//www.biomart.org
• NetAffx http//www.affymetrix.com

11
Pathway Analysis Tools

• PathVisio
• BioRAG
• MetaCore (GeneGO)
• Pathway-Express
• GenMAPP / MAPPFinder

12
PathVisio
www.pathvisio.org
13
Data Analysis with PathVisio
14
Pathway Analysis Workflow

1. Prepare your data
2. Import your data in PathVisio
3. Find enriched pathways
4. Visualize data on pathways
5. Export pathway images

15
1. Prepare your data
16
File Format

• PathVisio accepts delimited text files
• Prepare and export from Excel

17
File Format

• Export from R

write.table(myTable, file txtFile, col.names
NA, sep "\t", quote FALSE, na "NaN")
18
Identifier Systems

• PathVisio accepts many identifier systems
• Probes
• Affymetrix, Illumina, Agilent,...
• Genes and Proteins
• Entrez Gene, Ensembl, UniProt, HUGO,...
• Metabolites
• ChEBI, HMDB, PubChem,...

19
2. Import your data
20
Import Expression Data
21
Gene Database
A pathway
Your data
Entrez Gene 5326 153 4357 65543 2094 90218
4357
??
ENS0002114
P4235
22
Gene Database

• Download from www.pathvisio.org/wiki/PathVisioDown
  load
• 32 species
• supported

23
Identifier and System Code
24
Exception File
Exceptions file
25
Pgex File

• Imported data is stored in a .pgex file
• Load an existing dataset

26
3. Find enriched pathways
27
Statistics
Unchanged gene Changed gene

• Question
• Does the small circle have a higher percentage of
  changed genes than the large circle?
• Is this difference significant?

28
Calculate Z-scores

• The Z-score can be used as a measure for how much
  a subset of genes is different from the rest
• r changed genes in Pathway
• n total genes in Pathway
• R changed genes
• N total genes Other enrichment calculation
  methods
• Ackermann M et al., A general modular
  framework for gene set enrichment analysis,
  BMC bioinformatics, 2009

29
Z-score

• The Z-score is a ranking method.
• High Z-score ? selection is very different from
  the rest of the dataset
• Z-score 0 ? selection is not different at all

30
Criteria
Define criterion and select pathway collection
criterion
collection
31
Z-score Calculation
r changed genes in Pathway n total genes
in Pathway
r
n
32
Z-score Calculation
33
4. Visualize your data
34
Create a Visualization
Add/Remove Visualizations
Activate visualization options
35
Color by Data Values
36
Color Set based on Criterion
37
Color Set based on Gradient
38
Visualizations

• Gradient based
• Fold-change
• Rule based
• Significant genes

39
Gradient based
40
Rule based
41
5. Export Pathways
42
Export Pathway

• Export to image formats

PNG
43
Any data associated to a gene, protein or
metabolite
44
PathVisio Team

• Maastricht University
• Martijn van Iersel
• Thomas Kelder
• Chris Evelo
• Gladstone Institute (San Francisco)
• Alexander Pico
• Kristina Hanspers
• Bruce Conklin
• Around the world
• Open Source Community

45
Afternoon Session
46
Afternoon Session

• Pathway Analysis of liver data set with PathVisio
• Find enriched pathways in a WikiPathways
  analysis collection for rats
• Create visualization and set the data in a
  biological context