PCA analysis with R

1
PCA analysis with R
Analysis of the Attitude dataset data(attitude) a
ttitude pcalt-princomp(attitude) plot(pca) biplot(p
ca)
2
Homework

• (Due Sept 27) Read 'Gene shaving' as a method
  for identifying distinct sets of genes with
  similar expression patterns. Hastie et al. Genome
  Biology 2000 1(2)http//genomebiology.com/2000/1/
  2/research/0003/
• (Team, due Oct 4) Construct a dataset from the
  colon cancer dataset with 5 genes and all
  samples. These 5 genes should satisfy the
  following criteria 1) with signals larger than
  log2(100) in all the samples, 2) with the largest
  variation across samples. Regarding the samples
  as data points, do a PCA analysis to the dataset.
  Plot the transformed samples in the space of the
  first two principle components. Comment on your
  results.

3
PCA, Case study
'Gene shaving' as a method for identifying
distinct sets of genes with similar expression
patterns. Hastie et al. Genome Biology 2000 1(2)
4
(No Transcript)
5
(No Transcript)
6
1 gene remains
7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
Issues in dimension reduction

• Feature selection what genes (features) are used
  to construct the original high-dimensional space
  containing the samples points?
• Genes with large variation across samples so
  sample points are farther apart from each other
• Genes related by biological aspects (e.g. the
  same cell-cycle phase or function group). Then
  samples will separate in a way to reflect their
  relevance to these biological aspects
• Start from all genes and iteratively keep genes
  that contribute more to the first few principle
  components