getPLM.es: getPLM.es

In metahdep: Hierarchical Dependence in Meta-Analysis

Description

Calculates effect size estimates for a single study, based on a probe-level model, in preparation for a meta-analysis. It returns an ES.obj object containing the result.

Usage

getPLM.es(abatch, trt1, trt2, covariates=NULL, dep.grp=NULL, 
          sub.gn=NULL, bg.norm=TRUE)

Arguments

abatch	An AffyBatch object containing the data of interest.
trt1	A vector (or list of vectors) of array indices for treatment level 1 (control). If more than one test of differential expression is to be performed (for multiple covariate levels, for example), this should be a list of vectors; each trt1 / trt2 vector pair defines a comparison of interest.
trt2	A vector (or list of vectors) of array indices for treatment level 2 (treatment). If more than one test of differential expression is to be performed (for multiple covariate levels, for example), this should be a list of vectors; each trt1 / trt2 vector pair defines a comparison of interest.
covariates	(optional) A data.frame object representing covariate differences, if any, among the comparisons defined by trt1 / trt2 vector pairs. This data.frame should have a named column for each covariate to be considered in the meta-analysis, regardless of whether the covariate takes on multiple values in the study represented by the abatch argument. This data.frame must have a row for each comparison of interest, as defined by the trt1 / trt2 vector pairs. Elements of this data.frame should be coded numerically.
dep.grp	(optional) A single numeric value representing the dependence group number assigned to the study. Studies from the same research team may be considered hierarchically dependendent and share the same value.
sub.gn	(optional) A vector of geneNames (probe set ID's); the probe-level model will only be fit for these probesets. If NULL (default), all probesets are used.
bg.norm	(optional) A logical value specifying whether or not to perform background correction and normalization before fitting the probe-level model.

Details

For some subset of probesets in a gene expression study, this function calculates the effect size estimates based on Bolstad's probe-level model (Bolstad 2004), as described in Hu et al. (2006). Only two-group comparisons (treatment vs. control, for example) are supported. This is done in preparation for a meta-analysis of multiple gene expression studies.

Value

An object of class ES.obj

Author(s)

John R. Stevens, Gabriel Nicholas

References

Bolstad B. M. (2004), Low-level Analysis of High-density Oligonucleotide Array Data: Background, Normalization and Summarization, PhD dissertation, U.C. Berkeley.

Hu P., Greenwood C.M.T., and Beyene J. (2006), Integrative Analysis of Gene Expression Data Including an Assessment of Pathway Enrichment for Predicting Prostate Cancer, Cancer Informatics 2006:2 289-300.

Stevens J.R. and Nicholas G. (2009), metahdep: Meta-analysis of hierarchically dependent gene expression studies, Bioinformatics, 25(19):2619-2620.

See also the metahdep package vignette.

Examples

###
###  See the metahdep package vignette for a full example 
###

metahdep documentation built on Nov. 8, 2020, 8:03 p.m.