version 1.1

DESCRIPTION

@@ -2,17 +2,18 @@ Package: svapls
 Type: Package
 Title: Surrogate variable analysis using partial least squares in a
         gene expression study.
-Version: 1.0
-Date: 2012-07-23
+Version: 1.1
+Date: 2013-03-12
 Author: Sutirtha Chakraborty, Somnath Datta and Susmita Datta
-Maintainer: Sutirtha Chakraborty <s0chak10@louisville.edu>
+Maintainer: Sutirtha Chakraborty <statistuta@gmail.com>
 Depends: R (>= 2.0), class, stats, pls
 Description: Accurate identification of genes that are truly
-        differentially expressed over two sample varieties, after
-        adjusting for hidden subject-specific effects of residual
-        heterogeneity.
+        differentially expressed between two varieties of samples,
+        after adjusting for the hidden subject-specific effects of
+        residual heterogeneity.
 License: GPL-3
 Collate: fitModel.R svpls.R hfp.R
-Packaged: 2012-09-24 20:28:43 UTC; SUTIRTHA
+Packaged: 2013-03-13 02:53:31 UTC; SUTIRTHA
+NeedsCompilation: no
 Repository: CRAN
-Date/Publication: 2012-09-25 17:07:59
+Date/Publication: 2013-03-13 17:41:21

MD5

@@ -1,13 +1,11 @@
-22b427be5ded6e676355ef104798d908 *DESCRIPTION
-8020920c147a56242b4276aed68b61e3 *NAMESPACE
-a5074d94ccf8cce03a5bc531f91999d0 *R/fitModel.R
-43fe41ebb6f2b5302809d25d3a4a5dff *R/hfp.R
-8beb2a0dc9346babd370127d7664ef09 *R/svpls.R
-ba13a5b0ce37b9c5b4e3445f6d41d508 *data/batch.dat.rda
-531531b6d7d614fbfa6400aaa7f5bbee *data/hidden_fac.dat.rda
-9856fb023050eadb7ba73ba949fab63f *man/batch.dat.Rd
-8372275fad517d8ccfb853d40586b3a7 *man/fitModel.Rd
-89a702f4be4052f9c71f79644c669835 *man/hfp.Rd
+1e48e886f01b56784ac0cba7dfb64519 *DESCRIPTION
+acbef90a3eebe2f8baf23da7175ce99a *NAMESPACE
+6239c02ae48508b1eec0f0576d136017 *R/fitModel.R
+6eca99de58e84deffcfc480a535a6efa *R/hfp.R
+c91129a96eeb7c1d353389d01915ebd2 *R/svpls.R
+3175666e8f260fa1044d2cce65af83d1 *data/hidden_fac.dat.rda
+3f4818485f3f0081d7384c056ca789e0 *man/fitModel.Rd
+4fd4079856d3e29d62847159d5ab416f *man/hfp.Rd
 10e23268e511aba52e7436aeaa27de04 *man/hidden_fac.dat.Rd
-3b4f5961141ed460346da7ba6936912b *man/svapls-package.Rd
-90515ea4941d1a54b1c0fefbdece08c0 *man/svpls.Rd
+2fbbc2ae2f861571df414ab5282740f0 *man/svapls-package.Rd
+ccc334e123a778cff516d50f2148d039 *man/svpls.Rd

NAMESPACE

@@ -1,8 +1,6 @@
 importFrom(stats, heatmap)
 import(pls)
+exportClasses(fitModel, svpls)
+exportMethods(print, summary)
 export(fitModel, svpls, hfp)
-S3method(print, fitModel)
-S3method(print, svpls)
-S3method(summary, fitModel)
-S3method(summary, svpls)
 

R/fitModel.R

@@ -1,5 +1,4 @@
-fitModel <-
-function(k1,k2,Y,n.surr){
+fitModel <- function(k1,k2,Y,n.surr){
 Y <- Y[,-1]
 G <- nrow(Y)
 n <- k1 + k2
@@ -14,9 +13,9 @@ if (n.surr == 0){
 
    G.hat <- V1.hat <- GV1.hat <- rep(0,G)
    for (i in 1:G){
-      G.hat[i] <- mean(Y[i,]) - mean(Y)
-      V1.hat[i] <- mean(Y[,1:k1]) - mean(Y) 
-      GV1.hat[i] <- mean(Y[i,1:k1]) - mean(Y[,1:k1])      
+   	   G.hat[i] <- mean(Y[i,]) - mean(Y)
+   	   V1.hat[i] <- mean(Y[,1:k1]) - mean(Y) 
+   	   GV1.hat[i] <- mean(Y[i,1:k1]) - mean(Y[,1:k1])      
    }
    mu.hat <- mean(Y)
    V2.hat <- -(k1/k2)*V1.hat
@@ -171,12 +170,18 @@ if (n.surr > 0){
 if (n.surr == 0) coef <- coef1
 if (n.surr != 0) coef <- coef2
 class(coef) <- c("fitModel","list","vector")
-
 return(coef)
 }
 
 ## new summary function S3
-summary.fitModel <- function(object, ... ){
+summary.fitModel <- function(object){
+cat("mu.hat: \n", object$mu.hat, "\n")
+cat("G.hat: \n")
+print(object$G.hat) 
+cat("\n")
+cat("V.hat: \n")
+print(object$V.hat)
+cat("\n")
 cat("GV.hat: \n")
 print(object$GV.hat)
 cat("\n")
@@ -186,6 +191,15 @@ cat("\n")
 cat("beta.hat: \n")
 print(object$beta.hat)
 cat("\n")
+cat("GZ1.hat: \n")
+print(object$GZ1.hat)
+cat("\n")
+cat("VZ1.hat: \n")
+print(object$VZ1.hat)
+cat("\n")
+cat("Vhat.GV.hat: \n")
+print(object$vhat.gvh)
+cat("\n")
 cat("MSE: \n")
 print(object$MSE)
 cat("\n")
@@ -195,18 +209,17 @@ cat("\n")
 }
 
 ## new print function S3
-print.fitModel <- function(x, ...){
+print.fitModel <- function(x){
 cat("Estimated coefficients of the surrogate variables: \n")
 print(x$beta.hat)
 cat("\n")
+cat("Estimated Gene-variety interaction effects: \n")
+print(x$GV.hat)
+cat("\n")
 cat("Estimated Mean Squared Error of the fitted model: \n")
 print(x$MSE)
 cat("\n")
 cat("AIC value of the fitted model: \n")
 print(x$AIC)
 cat("\n")
-}
-
-
-
-
+}

R/hfp.R

@@ -5,6 +5,6 @@ function (obj,gen,ind,Y){
     gen.index <- rep(0,num)
     for (i in 1:num) gen.index[i] <- which(genes == gen[i]) 
     W_imp <- obj$PLS.imp
-    if (max(W_imp) != 0) heatmap(t(W_imp[gen.index,ind]),Rowv = NA,Colv = NA,labRow=as.character(ind),labCol=gen,xlab="Genes",ylab="Specified subjects")
+    if (max(W_imp) != 0) heatmap(t(W_imp[gen.index,ind]),Rowv = NA,Colv = NA,labRow=as.character(ind),labCol=gen,xlab="Specified subjects")
     if (max(W_imp) == 0) print("No trace of hidden subject-specific variation is found in the data")
 }

R/svpls.R

@@ -1,5 +1,4 @@
-svpls <-
-function(k1,k2,Y,pmax=3,fdr=0.05){
+svpls <- function(k1,k2,Y,pmax=3,fdr=0.05){
    G <- nrow(Y)
    k <- k1 + k2
    n <- G*k
@@ -12,14 +11,14 @@ function(k1,k2,Y,pmax=3,fdr=0.05){
    opt_model <- which(model_aic == min(model_aic))
 
 if(opt_model == 1){
-beta_hat <- NULL
-Y.corr <- PLS.imp <- matrix(0,500,20)
-gv_diff <- (1+k1/k2)*model_res[[1]][[4]][,1]
-mse <- model_res[[1]][[5]]
-aic.opt <- model_res[[1]][[6]]
-t <- pval.opt <- rep(0,G)
-for (i in 1:G){
- t[i] <- gv_diff[i]/sqrt((1-1/G)*(1/k1 + 1/k2)*mse)
+	beta_hat <- NULL
+	PLS.imp <- matrix(0,G,k)
+	gv_diff <- (1+k1/k2)*model_res[[1]][[4]][,1]
+	mse <- model_res[[1]][[5]]
+	aic.opt <- model_res[[1]][[6]]
+	t <- pval.opt <- rep(0,G)
+	for (i in 1:G){
+ 		t[i] <- gv_diff[i]/sqrt((1-1/G)*(1/k1 + 1/k2)*mse)
             pval.opt[i] <- 2*(1 - pt(abs(t[i]),n-2*G))            
       } 
       p.opt <- sort(pval.opt)
@@ -52,7 +51,7 @@ if (opt_model != 1){
     surr <- rep(0,k)
     for (l in 1:(opt_model-1)) surr <- surr + beta_hat[l]*sc[,l]
     for (i in 1:G) R[i,] <- surr   
-
+	 
     GZ.Z <- outer(est_gz1,sc[,1],function(u,v) u*v)
 
     for (i in 1:G){
@@ -65,7 +64,7 @@ if (opt_model != 1){
     t <- pval.opt <- rep(0,G)
     for (i in 1:G) {
          t[i] <- gv_diff[i]/sqrt(gv.diff_var[i] * mse.opt)
-         pval.opt[i] <- 2 * (1 - pt(abs(t[i]), n*G - 3*G - pmax))
+         pval.opt[i] <- 2 * (1 - pt(abs(t[i]), n - 3*G - pmax))
     }
     p.opt <- sort(pval.opt)
     sig <- 0
@@ -83,13 +82,13 @@ if (opt_model != 1){
 
 Y.corr <- cbind(Y[,1],Y[,-1] - PLS.imp)
 
-res <- list(opt.model=opt_model, PLS.imp=PLS.imp, Y.corr=Y.corr, pvalues.adj=pval.opt, genes = sig_genes.opt, AIC.opt = aic.opt)
+res <- list(opt.model=opt_model, PLS.imp=PLS.imp, Y.corr, pvalues.adj=pval.opt, genes = sig_genes.opt, AIC.opt = aic.opt)
 class(res) <- c("svpls","list","vector")
 return(res)
 }
 
 ## new summary function S3
-summary.svpls <- function(object, ...){
+summary.svpls <- function(object){
 cat("Optimum model is \n")
 print(object$opt_model)
 cat("AIC value for the optimal model is \n")
@@ -105,16 +104,11 @@ print(object$genes)
 }
 
 ## new print function S3
-print.svpls <- function(x, ... ){
+print.svpls <- function(x){
 cat("The Optimal Model is: \n")
 print(x$opt.model)
 cat("\n")
 cat("The differentially expressed genes are: \n")
 print(x$genes)
 cat("\n")
-}
-
-
-
-
-
+}

man/fitModel.Rd

@@ -1,7 +1,5 @@
 \name{fitModel}
 \alias{fitModel}
-\alias{summary.fitModel}
-\alias{print.fitModel}
 \title{
 Function to fit an ANCOVA model to the log transformed gene expression
 data, with a certain specified number of surrogate variables. 
@@ -38,9 +36,6 @@ fitModel(k1, k2, Y, n.surr)
 }
 }
 \value{
-\code{fitModel} returns an object of \code{\link{class}} "\code{fitModel}". The object 
-contains a list of the following components:
-
   \item{mu.hat }{Intercept (general mean effect).}
   \item{G.hat }{Main effects of the genes.}
   \item{V.hat }{Main effects of the varieties.}
@@ -57,8 +52,8 @@ contains a list of the following components:
 
 \references{
 Sutirtha Chakraborty, Somnath Datta and Susmita Datta. (2012) 
-Surrogate Variable Analysis Using Partial Least Squares in Gene 
-Expression Studies. Bioinformatics.
+Surrogate Variable Analysis Using Partial Least Squares (SVA-PLS) in Gene Expression Studies. Bioinformatics, 28(6): 799-806.
+
 Martens, H., Naes, T. (1989) Multivariate Calibration. Chicestor:Wiley.
 }
 
@@ -70,14 +65,15 @@ Sutirtha Chakraborty, Somnath Datta and Susmita Datta.
 \code{\link{svpls}}, \code{\link{hfp}}
 }
 \examples{
-data(batch.dat)
+data(hidden_fac.dat)
 
 ## Fitting an ANCOVA model with 5 surrogate variables
-fit <- fitModel(10,10,batch.dat,n.surr = 5)
+fit <- fitModel(10,10,hidden_fac.dat,n.surr = 5)
 print(fit)
 
 }
 
+\keyword{models}
 \keyword{classes}
 \keyword{print}
 \keyword{methods}

man/hfp.Rd

@@ -37,8 +37,7 @@ factors in the gene expression data.
 }
 \references{
 Sutirtha Chakraborty, Somnath Datta and Susmita Datta. (2012) 
-Surrogate Variable Analysis Using Partial Least Squares in Gene 
-Expression Studies. Bioinformatics.
+Surrogate Variable Analysis Using Partial Least Squares (SVA-PLS) in Gene Expression Studies. Bioinformatics, 28(6): 799-806.
 }
 \author{
 Sutirtha Chakraborty, Somnath Datta and Susmita Datta.
@@ -57,6 +56,7 @@ genes <- c(1,20,55,70,100,150,250,450)
 subjects <- c(1,4,7,10,11,15,17,20)
 
 hfp(fit,genes,subjects,hidden_fac.dat)
+
 }
 
 \keyword{print}

man/svapls-package.Rd

@@ -16,8 +16,8 @@ in a gene expression data.
 \tabular{ll}{
 Package: \tab svapls\cr
 Type: \tab Package\cr
-Version: \tab 1.0\cr
-Date: \tab 2012-07-23\cr
+Version: \tab 1.1\cr
+Date: \tab 2013-03-12\cr
 License: \tab GPL-3\cr
 }
 The package can be used to find the genes that are truly differentially expressed
@@ -35,19 +35,17 @@ specific factors, for any specified set of genes and subjects.
 \author{
 Sutirtha Chakraborty, Somnath Datta and Susmita Datta.
 
-Maintainer: Sutirtha Chakraborty <s0chak10@louisville.edu>
+Maintainer: Sutirtha Chakraborty <statistuta@gmail.com>
 }
 \references{
-Sutirtha Chakraborty, Somnath Datta and Susmita Datta. (2012) 
-Surrogate Variable Analysis Using Partial Least Squares in Gene 
-Expression Studies. Bioinformatics.
+Sutirtha Chakraborty, Somnath Datta and Susmita Datta. (SVA-PLS) in Gene Expression Studies. Bioinformatics 28(6):799-806.
 }
 
-\keyword{package}
+\keyword{models}
 
 \examples{
-data(batch.dat)
-fit <- svpls(10,10,batch.dat,pmax = 5)
+data(hidden_fac.dat)
+fit <- svpls(10,10,hidden_fac.dat,pmax = 5)
 fit$genes
 Y.corrected <- fit$Y.corr