Microarray Data Analysis

1
Microarray Data Analysis

• Statistics 5810/6810
• Fall 2004
• MWF 230, ENGR 204
• Adele Cutler
• adele_at_math.usu.edu
• Credits Many thanks to the the R core team and
  the Bioconductor developers, in particular
• Sandrine Dudoit
• Robert Gentleman
• Duncan Temple Lang

2
What is a Microarray?

• There are two main types
• spotted oligonucleotide arrays (e.g. Affy)
• two-color cDNA arrays
• An animation of the latter is provided at
• www.bio.davidson.edu/courses/genomics/chip/chip.ht
  ml

3
Why are we interested in Microarrays?

• Biologists
• Use genome-wide measures of expression (mRNA
  transcription levels) to predict biological
  attributes
• Identify gene interactions
• Compare genome-wide measures of expression
  (phylogenetics, etc )
• Lots more

4
Why are we interested in Microarrays? (ctnd)

• Computer Scientists
• Image processing (spot detection)
• Machine learning, neural nets
• e.g. how can we predict whether someone has a
  given disease (given 16,000 expression measures
  and a training dataset of 80 cases)?
• visualization
• Data warehousing and database management
• Lots more

5
Why are we interested in Microarrays? (ctnd)

• Statisticians
• large p small n problem (more variables than
  cases)
• Multiple comparisons
• Clustering and classification
• e.g. how can we learn about which genes are
  important in predicting whether a person has a
  given disease or not (given 16,000 expression
  measures and a training dataset of 80 cases)?
• visualization

6
What software will we use for microarray data
analysis?

• R for general data analysis www.r-project.org
• environment for statistical computing and
  graphics
• an open-source implementation of the S language
  (commercial Splus)
• CRAN www.cran.r-project.org
• omega www.omegahat.org
• Bioconductor for specialized microarray analysis
• www.bioconductor.org
• ImageJ for image processing/spot detection
• http//rsb.info.nih.gov/ij/

7
Installing R

• Available as source code, but easier to use
  precompiled versions for
• Windows
• MacOS
• Linux
• The R FAQs are useful if you run into problems

8
Installing R Details for Windows

• Go to www.r-project.org
• Under Download (l.h.s) select CRAN
• Select a mirror
• Save the windows version
• Select the appropriate operating system (Windows
  ?)
• Select base
• Select rw1091.exe
• Save, then execute to install

9
Installing bioconductor

• In R, do
• gtSys.putenv("http_proxy""http//proxy.usu.edu80/
  ")
• gtSys.getenv("http_proxy")
• gtsource("http//www.bioconductor.org/getBioC.R")
• gtgetBioC()
• The first two lines are only necessary if you
  encounter problems with the firewall. They must
  be used at the beginning of a new R session
  (before trying to access the web)