COMPARATIVE EVALUATION OF STATISTICAL TESTS FOR THE DETECTION OF DIFFERENTIALLY EXPRESSED GENES IN MULTIPLE TAG SAMPLING EXPERIMENTS
Chiara Romualdi, Stefania Bortoluzzi and
Gian Antonio Danieli
Departement of Biology and *CRIBI Biotecnology Centre, University of Padova, Italy
The comparison of several statistical methods currently used for detection of differentially expressed genes was attempted both by a simulation approach and by the analysis of data sets of human ESTs, obtained from UniGene.
Different statistical tests were used:the Audic and Claverie pairwise test (1997); the Fisher's exact test used by CGAP (O’Brien, 1997); Greller and Tobin test (1999); the R test (Stekel et al., 2000) and pairwise and general Chi-square tests.
A dedicated software was written (download IDEG.6).
The analysis of simulated data mimicking reality, the general Chi2 resulted the most efficient test, especially when dealing with variations affecting weakly expressed genes. This result was confirmed by the analysis of real sets of UniGene data. Moreover, this study showed that the Audic & Claverie’s method is the most efficient for detecting differences in gene expression when dealing with pairwise comparisons.
The analysis of UniGene-based dataset concerning human adult kidney as compared with two human kidney tumours , succeeded in identifying three novel genes, which appeared overexpressed in these tumours.
Further details about test statistics and data simulation are available.
GetProfiles Page generated by Chiara Romualdi firstname.lastname@example.org and Stefania Bortoluzzi email@example.com (May 15, 2001)