LmFit-Weights.R

From Organic Design wiki
Revision as of 23:06, 2 August 2006 by Sven (talk | contribs) (altering weights directly on MAList)
  1. ========================================================================
  2. This script examines the effect of weights and NA's on sigma estimates
  3. ========================================================================

library(limma) options(digits=2) packageDescription("limma", field="Version")

set.seed(2) n <- 50 preps <- 10

RG <- new("RGList") RG$R <- matrix(rnorm(n*preps,8,2), nc=preps) RG$G <- matrix(rnorm(n*preps,8,2), nc=preps) RG$printer <- structure(list(ngrid.c=1, ngrid.r=1, nspot.c=2, nspot.r=2), 

                        class="printlayout")

MA <- MA.RG(RG) colnames(MA$M) <- paste("array", 1:preps, sep="") rownames(MA$M) <- paste("gene", 1:n, sep="")

  1. ========================================================================
  2. 1) Analysis, no weights, no NA's
  3. ========================================================================

design <- rep(c(1,-1), preps/2) fit <- lmFit(MA, design=design, weights=NULL) fit <- eBayes(fit) topTable(fit, adjust="none")

  1. Sigma component - sd(x)

fit$sigma apply(MA$M * outer(rep(1, n), design), 1, sd)

  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled) * fit$sigma apply(MA$M * outer(rep(1, n), design), 1, function(x){sd(x)/sqrt(length(x))})


  1. ========================================================================
  2. 2) Analysis, adding some weights to MAList
  3. ========================================================================

MA$weights <- matrix(1, n,preps) MA$weights[row(MA$weights)<col(MA$weights)] <- 0 MA$weights

design <- rep(c(1,-1), preps/2) fit <- lmFit(MA, design=design, weights=MA$weights) fit <- eBayes(fit) topTable(fit, adjust="none") fit$df.residual

  1. Zero weights should be identical to NA's

MA$M[MA$weights==0] <- NA designMat <- outer(rep(1, n), design)

  1. Sigma component - sd(x)

fit$sigma apply(MA$M * designMat, 1, sd, na.rm=TRUE)

  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled) * fit$sigma apply(MA$M * designMat, 1, function(x,...){sd(x,...)/sqrt(length(x[!is.na(x)]))}, na.rm=TRUE)


  1. ========================================================================
  2. 3) Added some NA's for MAList (several lines above)
  3. ========================================================================

design <- rep(c(1,-1), preps/2) fit <- lmFit(MA, design=design, weights=MA$weights) fit <- eBayes(fit) topTable(fit, adjust="none") fit$df.residual

  1. Sigma component - sd(x)

fit$sigma apply(MA$M * designMat, 1, sd, na.rm=TRUE)

  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled) * fit$sigma apply(MA$M * designMat, 1, function(x,...){sd(x,...)/sqrt(length(x[!is.na(x)]))}, na.rm=TRUE)

  1. ========================================================================
  2. 3) Added weights and NA's simultaneously at random for MAList
  3. ========================================================================

set.seed(2) n <- 50 preps <- 10

RG <- new("RGList") RG$R <- matrix(rnorm(n*preps,8,2), nc=preps) RG$G <- matrix(rnorm(n*preps,8,2), nc=preps) RG$printer <- structure(list(ngrid.c=1, ngrid.r=1, nspot.c=2, nspot.r=2), 

                        class="printlayout")

MA <- MA.RG(RG)

ncoords <- sample(n,round(n/2)) pcoords <- sample(preps,round(n/2), replace=TRUE)

MA$M[cbind(ncoords, pcoords)] <- NA

ncoords <- sample(n,round(n/2)) pcoords <- sample(preps,round(n/2), replace=TRUE)

MA$weights <- matrix(1, nc=preps, nr=n) MA$weights[cbind(ncoords, pcoords)] <- 0

design <- rep(c(1,-1), preps/2) fit <- lmFit(MA, design=design, weights=MA$weights) fit$coeff fit <- eBayes(fit) topTable(fit, adjust="none") fit$df.residual

  1. Zero weights should be identical to NA's

MA$M[MA$weights==0] <- NA

  1. Sigma component - sd(x)

fit$sigma apply(MA$M * designMat, 1, sd, na.rm=TRUE)

  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled) * fit$sigma apply(MA$M * designMat, 1, function(x,...){sd(x,...)/sqrt(length(x[!is.na(x)]))}, na.rm=TRUE)


  1. ========================================================================
  2. Two groups common reference style analysis
  3. ========================================================================

RG <- new("RGList") RG$R <- matrix(rnorm(n*preps,8,2), nc=preps) RG$G <- matrix(rnorm(n*preps,8,2), nc=preps) RG$printer <- structure(list(ngrid.c=1, ngrid.r=1, nspot.c=2, nspot.r=2), 

                        class="printlayout")
  1. Ordering genes same ass topTable for convenience
  2. RG <- RG[c(4,2,1,3),]

MA <- MA.RG(RG) colnames(MA$M) <- paste("array", 1:preps, sep="") rownames(MA$M) <- paste("gene", 1:n, sep="")

design <- cbind(1, rep(c(0,1), each=5)) colnames(design) <- c("A","B-A") fit <- lmFit(MA, design=design, weights=NULL) fit <- eBayes(fit) topTable(fit, coef="B-A", adjust="none")


  1. Sigma component - sd(x). The design matrix parametrisation is fitting a more complex
  2. model, decreasing the rank by 1 which gives slightly modified estimates of sigma
  3. The code in lm.series which calculates fit$sigma is;
  1. fit$sigma <- sqrt(colMeans(fit$effects[(fit$rank + 1):narrays, , drop = FALSE]^2))

fit$sigma apply(MA$M , 1, sd, na.rm=TRUE) # Not quite the same

plot(fit$sigma, apply(MA$M , 1, sd, na.rm=TRUE))

  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled[,2]) * fit$sigma # Check differences here! apply(MA$M, 1, function(x,...){sd(x,...)/sqrt(length(x[!is.na(x)]))}, na.rm=TRUE)

a <- drop(fit$stdev.unscaled[,2]) * fit$sigma # Check differences here! b <- apply(MA$M, 1, function(x,...){sd(x,...)/sqrt(length(x[!is.na(x)]))}, na.rm=TRUE)

plot(a, b*2) abline(0,1)

  1. ========================================================================
  2. Two groups common reference style analysis (alternate parametrisation)
  3. ========================================================================

design <- cbind(rep(c(1,0), each=5), rep(c(0,1), each=5)) colnames(design) <- c("A","B") cont.matrix <- makeContrasts("B-A" = B-A, levels=design)

fit2 <- lmFit(MA, design=design, weights=NULL) fit2 <- contrasts.fit(fit2, cont.matrix) fit2 <- eBayes(fit2) topTable(fit2, adjust="none")

fit$sigma # identical fit2$sigma

  1. ========================================================================
  2. Affy style analysis (using R, G channels here)
  3. ========================================================================

library(convert) RG <- new("RGList") RG$R <- matrix(rnorm(n*preps,8,2), nc=preps) RG$G <- matrix(rnorm(n*preps,8,2), nc=preps) RG$printer <- structure(list(ngrid.c=1, ngrid.r=1, nspot.c=2, nspot.r=2), 

                        class="printlayout")

exprSet <- as(RG, "exprSet")

design <- cbind(rep(1, 20), rep(c(1,0), 10)) colnames(design) <- c("A","B-A")

fit <- lmFit(exprSet, design=design) fit <- eBayes(fit) topTable(fit, adjust="none", coef="B-A") fit$df.residual fit$sigma

design <- cbind(rep(c(0,1), 10), rep(c(1,0), 10)) colnames(design) <- c("A","B") cont.matrix <- makeContrasts("B-A" = B-A, levels=design)

fit <- lmFit(exprSet, design=design) fit <- contrasts.fit(fit, cont.matrix) fit <- eBayes(fit) topTable(fit, adjust="none") fit$df.residual fit$sigma


  1. Sigma component - sd(x)

fit$sigma apply(exprs(exprSet), 1, sd) plot(fit$sigma, apply(exprs(exprSet), 1, sd)) abline(0,1)

  1. apply(exprs(exprSet) * outer(rep(1, n), (design[,2]*2)-1), 1, sd)
  2. apply(exprs(exprSet) * outer(rep(1, n), design), 1, sd)
  1. Stderr component - se(x_bar)

drop(fit$stdev.unscaled) * fit$sigma apply(MA$M * outer(rep(1, n), design), 1, function(x){sd(x)/sqrt(length(x))})

Retrieved from "https://organicdesign.nz/wiki/index.php?title=LmFit-Weights.R&oldid=26009"