Learn R Programming

plw (version 1.30.0)

lmw: Locally Moderated Weighted-t.

Description

Computes Locally Moderated Weighted t-test for microarray data.

Usage

lmw(x,design=rep(1,ncol(x)),contrast=matrix(1), meanX=NULL, maxIter = 200, epsilon = 1e-06, verbose = TRUE, nknots = 10, nOut = 2000, nIn = 4000, knots = NULL, checkRegulation = TRUE)

Arguments

x
Data, log2 expression indexes.
design
design matrix
contrast
contrast matrix
meanX
Covariate used to model scale parameter, default=NULL (see details)
maxIter
maximum number of iterations
epsilon
convergence criteria
verbose
print computation info or not
nknots
Number of knots of spline for $v$
nOut
Parameter for calculating knots, see getKnots
nIn
Parameter for calculating knots, see getKnots
knots
Knots, if not NULL it overrides nknots, nOut and nIn
checkRegulation
If TRUE, data is checked for a correct specified contrast (see details)

Value

Sigma
Estimated covariance matrix for $y=P' x$
m
Estimated shape parameter for inverse-gamma prior for gene variances
v
Estimated scale parameter curve for inverse-gamma prior for gene variances
converged
T if the EM algorithms converged
iter
Number of iterations
modS2
Moderated estimator of gene-specific variances
histLogS2
Histogram of log(s2) where s2 is the ordinary variance estimator
fittedDensityLogS2
The fitted density for log(s2)
logs2
Variance estimators, logged with base 2.
t
Moderated t-statistic
coefficients
Estimated contrast
p.value
P-value from the moderated t-statistic
dfT
Degrees of freedom of the moderated t-statistic
weights
Weights for estimating the contrast
P
Transformation matrix
beta
Estimated parameter vector $beta$ of spline for $v(x)$
knots
The knots used in spline for $v(x)$

Details

This function computes the Locally Moderated Weighted-t statistic (LMW) described in $A$strand (2007b), thus calculating locally moderated weighted t-statistic, p-value and log2(FC) for each row of the data matrix x.

Each gene g (row of x) is modeled as: $$y_g|c_g \sim N(\mu_g,c_g\Sigma)$$ $$c_g \sim \mbox{InvGamma}(m/2,m\nu/2)$$ where $v$ is function of the mean intensity: $v(mean(mu_g))$, $N$ denotes a multivariate normal distribution, $Sigma$ is a covariance matrix and $InvGamma(a,b)$ is the inverse-gamma distribution with density function $$f(x)=(b)^{a} \exp\{-b/x\} x^{-a-1}/\Gamma(a)$$

Given the design matrix D, $mu_g$ equals $D*gamma_g$, and given the contrast matrix C the hypothesis $C*gamma_g=0$ is tested. C should be a one row matrix of same length as the column vector $gamma_g$.

See examples on how to specify the design and contrast matrices.

A cubic spline is used to parameterize the smooth function $v(x)$ $$\nu(x) = \exp\{ H(x)^T \beta \}$$ where $H:R->R^(2p-1)$ is a set B-spline basis functions for a given set of p interior spline-knots, see chapter 5 of Hastie et al. (2001).

For details about the model see Kristiansson et al. (2005), $A$strand et al. (2007a,2007b).

As specified above, $v$ is modeled as a function of mean intensity: $v(mean(mu_g))$. If the parameter meanX is not NULL, meanX is used instead of the mean intensity when modeling $v$. Thus, if meanX is not NULL, meanX must be a vector of length equal to the number of rows of the data matrix x.

The parameter estimation procedure is based on the assumption that the specified contrast is close to zero for most genes, or at least that the median contrast over all genes is close to zero. A check is run on data to validate this assumptions. If the checking fails, with the error message "warning: most genes appears to be regulated..." and if YOU ARE SURE that the design and contrast is correct, use checkRegulation=FALSE.

References

Hastie, T., Tibshirani, R., and Friedman, J. (2001). The Elements of Statistical Learning, volume 1. Springer, first edition.

Kristiansson, E., Sj$o$gren, A., Rudemo, M., Nerman, O. (2005). Weighted Analysis of Paired Microarray Experiments. Statistical Applications in Genetics and Molecular Biology 4(1)

$A$strand, M. et al. (2007a). Improved covariance matrix estimators for weighted analysis of microarray data. Journal of Computational Biology, Accepted.strand

$A$strand, M. et al. (2007b). Empirical Bayes models for multiple-probe type arrays at the probe level. Bioinformatics, Submitted 1 October 2007.

See Also

estimateSigmaMVbeta, plw

Examples

Run this code
# ------------------------------------------
# Example analyzing the 6 arrays in the 
# AffySpikeU95Subset data set

# Loading the data
data(AffySpikeU95Subset)

# Defining design and contrast matrix
group<-factor(rep(1:2,each=3))
design<-model.matrix(~group-1)
contrast<-matrix(c(1,-1),1,2)

# Computing RMA expression index
data.rma<-exprs(rma(AffySpikeU95Subset))

# Analyzing
model1<-lmw(data.rma,design=design,contrast=contrast,epsilon=0.01)

## Look at fitted vs observed density for log(s2)
varHistPlot(model1)

## Look at fitted curve for scale parameter
scaleParameterPlot(model1)

Run the code above in your browser using DataLab