SimAffyDat {SharedHT2} | R Documentation |
The package SharedHT2
contains one dataset SimAffyDat
for testing purposes. The dataframe has 12625 rows and 6 columns.
The rows are named after the hgu95 affy gene identifiers in order
that the links to the GeneCard database work. The six column variables
represent hypothetical base two logged expression ratios for
three replicates of treatment 1 versus control and treatment 2
versus control.
data(SimAfyDat)
All but the first 100 rows have population mean zero. The first 100 rows have group specific population means equal to $7.5/3^{0.5}$ times the square root of the averaged sample variances across genes. Strikingly, eye inspection of the dataset (compare the first 100 rows to the rows which follow) reveals positive and negative numbers among observations representing all but the first 100 rows. However, all tests except the emprical Bayes Hotelling T-squared tests are unable to correctly identify these "genes" without high false positive rates. The sampling distribution used in the simulation is the Wishart/Inverse Wishart model. See the manuscript "sharedHT2.pdf" in the ./doc directory.
Grant Izmirlian izmirlian@nih.gov
Izmirlian, G and Xu, J.-L. (2002), The Shrinkage Variance Hotelling T-Squared Test for Genomic Profiling Studies, NCI technical report.
EB.Anova
, EBfit
, TopGenes
,
SimW.IW
, Simnu.mix