est.lme: Estimate order restricted variance components for a two way...

In orQA: Order Restricted Assessment Of Microarray Titration Experiments

Description

Estimates variance components of a two way cross classification mixed model with an order restricted fixed effect, a random effect and random interaction term.

Usage

est.lme(y,ia,ib)

Arguments

y	a numeric matrix, for the lines of which univariate variance decompositions should be estimated
ia	integer vector of length ncol(y) specifying the ordered levels of the order restricted fixed effect as integers between 1 and n where n is the number of levels
ib	integer vector or factor specifying the levels of the random term (e.g. individuals)

Details

est.lme estimates the variance components for a two way cross classification mixed model with random interaction. The order restriction on the fixed effect used to improve the estimates by pooling levels of the fixed effect using isotonic regression according to the observed order in the measurements. Estimation is then done using functionality provided by the package <nlme>.

Value

sb	estimates of the random effect variance component
sg	estimates of the interaction term variance component
se	estimates of the residual error variance
tss	total sum of squares (useful to obtain normalized scales between estimates from different measurements)

Author(s)

Florian Klinglmueller <float_at_lefant.net>

References

Klinglmueller, F., Tuechler, T., Posch, M. (2010) "Cross Platform Comparison Of Microarray Data Using Order Restricted Inference" Under Review

Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D.,R Development Core Team (2010) "nlme: Linear and Nonlinear Mixed Effects Models"

Barlow, R. E., Bartholomew, D. J., Bremner, J. M., and Brunk, H. D. (1972) "Statistical inference under order restrictions"; Wiley, London.

Robertson, T., Wright,F. T. and Dykstra, R. L. (1988) "Order Restricted Statistical Inference"; Wiley, New York.

Examples

g <- rep(1:4,each=10)
r <- rep(rep(1:2,each=5),4)

## No differences (global null)
nulldata <- matrix(rnorm(400),nc=40)

## estimation
res <- est.lme(nulldata,g,r)
round(apply(res,2,summary),2)

orQA documentation built on May 1, 2019, 10:31 p.m.