Cargando…

The [Image: see text]-Version of the Cramér-von Mises Test for Two-Sample Comparisons in Microarray Data Analysis

Distribution-free statistical tests offer clear advantages in situations where the exact unadjusted [Image: see text]-values are required as input for multiple testing procedures. Such situations prevail when testing for differential expression of genes in microarray studies. The Cramér-von Mises tw...

Descripción completa

Detalles Bibliográficos
Autores principales: Xiao, Yuanhui, Gordon, Alexander, Yakovlev, Andrei
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Springer 2006
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3171322/
https://www.ncbi.nlm.nih.gov/pubmed/18427586
http://dx.doi.org/10.1155/BSB/2006/85769
Descripción
Sumario:Distribution-free statistical tests offer clear advantages in situations where the exact unadjusted [Image: see text]-values are required as input for multiple testing procedures. Such situations prevail when testing for differential expression of genes in microarray studies. The Cramér-von Mises two-sample test, based on a certain [Image: see text]-distance between two empirical distribution functions, is a distribution-free test that has proven itself as a good choice. A numerical algorithm is available for computing quantiles of the sampling distribution of the Cramér-von Mises test statistic in finite samples. However, the computation is very time- and space-consuming. An [Image: see text] counterpart of the Cramér-von Mises test represents an appealing alternative. In this work, we present an efficient algorithm for computing exact quantiles of the [Image: see text]-distance test statistic. The performance and power of the [Image: see text]-distance test are compared with those of the Cramér-von Mises and two other classical tests, using both simulated data and a large set of microarray data on childhood leukemia. The [Image: see text]-distance test appears to be nearly as powerful as its [Image: see text] counterpart. The lower computational intensity of the [Image: see text]-distance test allows computation of exact quantiles of the null distribution for larger sample sizes than is possible for the Cramér-von Mises test.