ANOVA tutorial

Supplement to Using ANOVA for gene selection from microarray studies of the nervous system.
Pavlidis P pubmed reprint

Abstract

Methods are presented for detecting differential expression using statistical hypothesis testing methods including analysis of variance (ANOVA). Practicalities of experimental design, power, and sample size are discussed. Methods for multiple testing correction and their application are described. Instructions for running typical analyses are given in the R programming environment. R code and the sample data set used to generate the examples are available here.

Contact

paul@chibi.ubc.ca

Old Supplement Link

http://chibi.ubc.ca/faculty/pavlidis/lab/aovmethods/

Supplement: Sample files and code

Sample data file. This contains data for 1000 probes from the data of Sandberg et al.(PNAS 2000).

 

R code. This file contains the commands used to execute the analysis, and some additional commentary. (Updated Nov 17 2010)

 

Stack overflow errors? We have found a a problem with running part of the code when adapted to large data sets having more genes (e.g., 10,000). This is due to a bug in R (at least as of 1.6.2) combined with our inelegant R code. Thanks to Bill Noble for discovering the problem and Tom Blackwell for the work-around. See the R code for the work-around explanation.

Supplement: Output files

The following files are created during the demo analysis. They will not appear properly if viewed in your web browser. Instead, they are meant to be downloaded and opened in Excel or another spreadsheet program.

 

Supplement: Color version of Figure 3