Bootstrapping of gene-expression data improves and controls the false discovery rate of differentially expressed genes
<p>Abstract</p> <p>The ordinary-, penalized-, and bootstrap <it>t</it>-test, least squares and best linear unbiased prediction were compared for their false discovery rates (FDR), <it>i.e. </it>the fraction of falsely discovered genes, which was empirically...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | deu |
| Published: |
BMC
2004-03-01
|
| Series: | Genetics Selection Evolution |
| Subjects: | |
| Online Access: | http://www.gsejournal.org/content/36/2/191 |
| Summary: | <p>Abstract</p> <p>The ordinary-, penalized-, and bootstrap <it>t</it>-test, least squares and best linear unbiased prediction were compared for their false discovery rates (FDR), <it>i.e. </it>the fraction of falsely discovered genes, which was empirically estimated in a duplicate of the data set. The bootstrap-<it>t</it>-test yielded up to 80% lower FDRs than the alternative statistics, and its FDR was always as good as or better than any of the alternatives. Generally, the predicted FDR from the bootstrapped <it>P</it>-values agreed well with their empirical estimates, except when the number of mRNA samples is smaller than 16. In a cancer data set, the bootstrap-<it>t</it>-test discovered 200 differentially regulated genes at a FDR of 2.6%, and in a knock-out gene expression experiment 10 genes were discovered at a FDR of 3.2%. It is argued that, in the case of microarray data, control of the FDR takes sufficient account of the multiple testing, whilst being less stringent than Bonferoni-type multiple testing corrections. Extensions of the bootstrap simulations to more complicated test-statistics are discussed.</p> |
|---|---|
| ISSN: | 0999-193X 1297-9686 |