Identify differentially expressed genes by combining effect sizes
Function to fit the meta-analytic fixed- and random-effects models.The data consists of effect sizes and corresponding variances from your own method/calculations.
MetaDE.ES(x, meta.method = c("FEM", "REM"))
- a list with components.
- ES: The observed effect sizes.
- Var: The observed variances corresponding to
- perm.ES: The effect sizes calculated from permutations,
perm.ESis NULL if the argum
- a character string specifying whether a fixed- or a random/mixed-effects model should be fitted. A fixed-effects model is fitted when using meta.method="FEM". Random-effects model is fitted by setting meta.method equal to "REM". See "Details".
The function can be used to combine any of the usual effect size used in meta-analysis,such as standardized mean differences.Simply specify the observed effect sizes via the x$ES
and the corresponding variances vis x$Var. If the effect sizes and corresponding varicances calculated from permutation are available,then specify them by x$perm.ES and x$perm.Var,
paired is a vecter of logical values to specify whethe the corresponding study is paired design or
not. If the study is pair-designed, the effect sizes (corresponding variances) are calcualted using the formula in morris's
paper, otherwise calculated using the formulas in choi et al.
In addition, if the components of x, perm.ES and perm.Var, are not "NULL", the p-values are calculated using permutation method, otherwise, the p-values are calculated using parametric method by
assupming the z-scores following a standard normal distribution.
- The object is a list containing the following components:
zval test statistics of the aggregated value. pval p-values for the test statistics. FDR A matrix with one column which has the corrected p-values using Benjamini and Hochberg method (see
Qval test statistics for the test of heterogeneity. Qpval p-values for the test of heterogeneity. tau2 estimated amount of (residual) heterogeneity.
Choi et al, Combining multiple microarray studies and modeling interstudy variation. Bioinformatics,2003, i84-i90. Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B, 57, 289:300.
#---example 1: Meta analysis of Differentially expressed genes between two classes----------# label1<-rep(0:1,each=5) label2<-rep(0:1,each=5) exp1<-cbind(matrix(rnorm(5*20),20,5),matrix(rnorm(5*20,2),20,5)) exp2<-cbind(matrix(rnorm(5*20),20,5),matrix(rnorm(5*20,1.5),20,5)) x<-list(list(exp1,label1),list(exp2,label2)) ind.res<-ind.cal.ES(x,paired=rep(FALSE,2),nperm=100) MetaDE.ES(ind.res,meta.method='REM')