A random variance model for detection of differential gene expression in small microarray experiments.

摘要: Motivation: Microarray techniques provide a valuable way of characterizing the molecular nature disease. Unfortunately expense and limited specimen availability often lead to studies with small sample sizes. This makes accurate estimation variability difficult, since variance estimates made on gene by basis will have few degrees freedom, assumption that all genes share equal is unlikely be true. Results: We propose model which within variances are drawn from an inverse gamma distribution, whose parameters estimated across genes. results in test statistic minor variation those used standard linear models. demonstrate assumptions valid experimental data, has more power than tests pick up large changes expression, while not increasing rate false positives. Availability: method incorporated into BRB-ArrayTools version 3.0 (http://linus.nci.nih.gov/BRB-ArrayTools.html). Supplementary material: ftp://linus.nci.nih.gov/pub/ techreport/RVM_supplement.pdf Contact: wrightge@mail.nih.gov