Pathways Analysis using Protein Expression Data

1
Pathways Analysis using Protein Expression Data

• Venkatesh Jitender (vjk_at_cbmi.upmc.edu)
• Dr. Vanathi Gopalakrishnan
• Center for Biomedical Informatics, UPMC

2
Bio-Chemical Pathways Protein Expression
Profiles

• Organisms function through intricate networks of
  chemical reactions and interacting molecules.
• Metabolic Pathway
• Can be thought of as a state representation
  network
• Regulatory Pathway
• Can be thought of as a switch activating or
  de-activating diagram
• Protein Expression Profiles
• Mass- spectrometry
• Protein Array technology, etc.

3
KEGG Database

• Suite of Databases and Associated Software
  http//www.genome.ad.jp/kegg/kegg2.htmlpathway
• Includes
• PATHWAY database
• GENES PROTEINS (Genes / SSD / KO databases)
• CHEMICAL COMPOUNDS REACTIONS (Compound /
  Reaction database)
• Statistics
• Number of pathways10,677(PATHWAY database)
• Number of reference pathways226(PATHWAY database)

4
Metabolic Pathway
5
Regulatory Pathway
6
Proposal

• Identify co- relation between observed protein
  expression profiles and Biochemical Pathways.
• Provide an online web-service to predict the
  Most-Probable pathway
• Rank putative pathways based on
• Activity score analyze difference in
  expression profiles
• Co regulation score analyze similarity in
  expression profiles within a pathway
• Cascade score analyze structure, measuring
  activity and coregulation

7
Project Architecture
Putative Pathway Set
Application Server
KEGG Server KEGG API
Analysis Routine

get_pathway_by_enzyme
Recommended Pathway
User submits sequence data
8
Analysis Routine
Application Server

• Analysis Routine
• Score Pathway based on
• Activity
• Coregulation
• Cascade

Suggested Pathway
Putative Pathway Set
9
Implementation Details

• Option of program being hosted public (server) /
  standalone application
• Application server JAVA servlets, JSP
• Analysis Routine C/ JAVA
• Messaging System HTTP
• Communication Protocol SOAP (xml format)

10
Applications

• Pathway knowledge helpful in
• Deeper understanding of cell mechanisms under
  various conditions
• Predict possible drug targets,
• Provide end- user on most probable pathway set,
  from a publicly available database