The fundamental goal of microarray experiments is to identify genes that are differentially expressed in the conditions being studied.  Comparison statistics can be used to help identify differentially expressed genes and cluster analysis can be used to identify patterns of gene expression and to segregate a subset of genes based on these patterns.

Normalization: Methods for removing systematic variation from microarray data

•  Global methods
•  Intensity dependent methods

Statistical Significance

•  Fold Change

  Although simple and intuitive, fold change does not address the reproducibility of the observed difference and cannot be used to determine the statistical significance.

•  Comparison statistics use the replicates to assign a confidence level as to whether the gene is differentially expressed.
  •  Parametric
    •  T-test
     
    •  Welch’s t-test
    •  ANOVA (See GeneSifter's July 20 2006 Webinar on this)
     
     
  •  Non-parametric
    •  Wilcoxon Rank Sum, Kruskal-Wallis, Permutation t-test

Comparison tests require replicates and use the variability within the replicates to assign a confidence level as to whether differences in gene are significance or due to chance.

Correction for multiple testing - Methods for adjusting the p-value from a comparison test based on the number of tests performed. These adjustments help to reduce the number of false positives in an experiment.

•  FWER – Adjusts the p-value so that it reflects the chance of at least 1 false positive being found in the list.
  •  Bonferonni, Holm, W & Y MaxT
• FDR – Adjusts the p-value so that it reflects the frequency of falso positives in the list.
  • Benjamini and Hochberg

FWER is more conservative, but the false discovery rate(FDR) is usually acceptable for “discovery” experiments.

Cluster Analysis: clustering methods are descriptive or exploratory tools that can be used to identify groups within complex datasets. They can be used to identify patterns of gene expression in microarray datasets.

•  Visualization: Methods such as hierarchical clustering can help identify patterns in a large dataset.
•  Partitioning: This type of cluster analysis can be used to separate data into discrete groups.
•  K-means
•  PAM
• Silhouettes

Cluster analysis can be used to identify patterns within large datasets and to partition genes based on these patterns.