Clustering%20Algorithms%20to%20make%20sense%20of%20Microarray%20data:%20Systems%20Analyses%20in%20Biology - PowerPoint PPT Presentation

About This Presentation
Title:

Clustering%20Algorithms%20to%20make%20sense%20of%20Microarray%20data:%20Systems%20Analyses%20in%20Biology

Description:

Beloit Wisconsin, June 2004. Biological Questions ... Download: version 1.4 (v2.2 has bug): http://rana.lbl.gov/EisenSoftware.htm. Load data ... – PowerPoint PPT presentation

Number of Views:48
Avg rating:3.0/5.0
Slides: 16
Provided by: brian631
Learn more at: http://bioquest.org
Category:

less

Transcript and Presenter's Notes

Title: Clustering%20Algorithms%20to%20make%20sense%20of%20Microarray%20data:%20Systems%20Analyses%20in%20Biology


1
Clustering Algorithms to make sense of Microarray
data Systems Analyses in Biology
  • Doug Welsh and Brian Davis
  • BioQuest Workshop
  • Beloit Wisconsin, June 2004

2
Biological Questions
  • What are the differences between cancer cells and
    normal cells?
  • What are the differences in gene expression
    between cancer cells and normal cells?
  • Can you guess at the cellular sub-systems that
    may be affected by cancer?
  • What are the cellular processes (pathways) that
    might differ between cancer cells and normal
    cells?
  • Can you guess at the components (proteins) of the
    pathways that might be involved in cancer

3
Goals
  • Systems Biology (shift focus among levels of
    knowledge)
  • Biology
  • Gene Expression (technique)
  • Pathways
  • DNA Replication
  • Individual Proteins
  • Math
  • Clustering Algorithms (theory and technique)
  • Statistics
  • Medicine (human phenotype)

4
Goals
Cluster
Programming
Medicine
Stats
Math
Biology
Cell Biology
Knowledge
Pathway
Physics
Optics
Robotics
Protein
5
Goals
Cluster
Programming
Medicine
Stats
Math
Biology
Cell Biology
Knowledge
Pathway
Physics
Tools
Optics
Robotics
Protein
6
Problem Space
Bedrock web site http//bioquest.org/bedrock/prob
lem_spaces/
7
Problem Space
DNA Replication Cell Cycle (Depends on Paper)
Microarray Files Gene Annotation
Microarray Analysis Pathway Analysis Statistical
Analysis
8
Assumptions
  • Assume co-expression of genes has significance.
  • We can generate A LOT of data.

Wheat and Chaff
  • Clustering algorithms and viewing software allow
    a researcher to focus on subsets of
    (significant) data at a time.

9
Project
  • Paper Singh D. et al. (2002) Gene expression
    correlates of clinical prostate cancer behavior.
    Cancer Cell Mar1(2)203-9.
  • Questions What is the testable hypothesis? How
    is it tested? What are the results? Are the
    conclusions valid?
  • Are there other (better?) ways to test this
    hypothesis? Are there better hypotheses to
    formulate?

10
Biological Questions
  • What are the differences between cancer cells and
    normal cells?
  • What are the differences in gene expression
    between cancer cells and normal cells?
  • Can you guess at the cellular sub-systems that
    may be affected by cancer?
  • What are the cellular processes (pathways) that
    might differ between cancer cells and normal
    cells?
  • Can you guess at the components (proteins) of the
    pathways that might be involved in cancer

11
Cluster
  • Download version 1.4 (v2.2 has bug)
  • http//rana.lbl.gov/EisenSoftware.htm
  • Load data
  • Data transform

12
TreeView
  • Download latest version
  • http//rana.lbl.gov/EisenSoftware.htm
  • Load data

13
Analysis
  • Clustering may reveal organizational units
  • What are these proteins and what processes are
    they involved in?

14
Next Steps
  • Hand off clusters of organizational units to Doug
    (GeneMAPP and MAPPFinder What are these
    proteins in the context of cellular pathways?)
  • Investigate interesting single proteins (e.g.,
    with NCBI tools).
  • Are these proteins conserved? (do yeast get
    cancer?)
  • What is the molecular basis of cancer?

15
Goals
Cluster
Medicine
Stats
Math
Biology
Cell Biology
Knowledge
Pathway
Physics
Optics
Robotics
Protein
Write a Comment
User Comments (0)
About PowerShow.com