Add examples, fix bugs

Russel88 · Mar 4, 2019 · 29e0fe4 · 29e0fe4
1 parent 9266758
commit 29e0fe4
Show file tree

Hide file tree

Showing 125 changed files with 1,364 additions and 71 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -7,7 +7,7 @@ Authors@R: person("Jakob", "Russel",
 Description: Runs multiple differential abundance/expressions methods,
     with the same standardized input and output, 
     to compare their results and test which are better at finding spiked features. 
-Depends: R (>= 3.5)
+Depends: R (>= 3.4)
 biocViews: Sequencing,
     RNASeq,
     Microbiome,
@@ -50,3 +50,4 @@ License: GPL (>= 3) | file LICENSE
 Encoding: UTF-8
 LazyData: true
 RoxygenNote: 6.1.1
+BugReports: https://github.com/Russel88/DAtest/issues
diff --git a/NAMESPACE b/NAMESPACE
@@ -64,7 +64,7 @@ export(norm_alr)
 export(norm_clr)
 export(powerDA)
 export(preDA)
-export(prune.tests.DA)
+export(pruneTests)
 export(runtimeDA)
 export(spikein)
 export(testDA)

diff --git a/NEWS b/NEWS
diff --git a/R/DA.adx.R b/R/DA.adx.R
@@ -5,6 +5,15 @@
 #' @param predictor The predictor of interest. Factor, OR if \code{data} is a \code{phyloseq} object the name of the variable in \code{sample_data(data)} in quotation
 #' @param ... Additional arguments for the \code{aldex} function
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running ALDEx2
+#' res <- DA.adx(data = mat, predictor = pred)
 #' @export
 
 DA.adx <- function(data, predictor, ...){

diff --git a/R/DA.aoa.R b/R/DA.aoa.R
@@ -11,6 +11,15 @@
 #' @param allResults If TRUE will return raw results from the \code{aov} function
 #' @param ... Additional arguments for the \code{aov} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running ANOVA on each feature
+#' res <- DA.aoa(data = mat, predictor = pred)
 #' @export
 
 DA.aoa <- function(data, predictor, covars = NULL, p.adj = "fdr", delta = 1, allResults = FALSE, ...){
@@ -43,7 +52,7 @@ DA.aoa <- function(data, predictor, covars = NULL, p.adj = "fdr", delta = 1, all
   }
 
   # Zero-correction
-  count_table <- apply(count_table, 2, function(y) vapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
+  count_table <- apply(count_table, 2, function(y) sapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
   if(any(count_table <= 0)) stop("Zero-correction failed. Dataset likely contains too many zeroes")
 
   # ALR transformation

diff --git a/R/DA.aoc.R b/R/DA.aoc.R
@@ -9,6 +9,15 @@
 #' @param allResults If TRUE will return raw results from the \code{aov} function
 #' @param ... Additional arguments for the \code{aov} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running ANOVA on each feature
+#' res <- DA.aoc(data = mat, predictor = pred)
 #' @export
 
 DA.aoc <- function(data, predictor, covars = NULL, p.adj = "fdr", delta = 1, allResults = FALSE, ...){
@@ -41,7 +50,7 @@ DA.aoc <- function(data, predictor, covars = NULL, p.adj = "fdr", delta = 1, all
   }
 
   # Zero-correction
-  count_table <- apply(count_table, 2, function(y) vapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
+  count_table <- apply(count_table, 2, function(y) sapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
   if(any(count_table <= 0)) stop("Zero-correction failed. Dataset likely contains too many zeroes")
 
   # ALR transformation

diff --git a/R/DA.aov.R b/R/DA.aov.R
@@ -9,6 +9,15 @@
 #' @param allResults If TRUE will return raw results from the \code{aov} function
 #' @param ... Additional arguments for the \code{aov} function
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running ANOVA on each feature
+#' res <- DA.aov(data = mat, predictor = pred)
 #' @export
 
 DA.aov <- function(data, predictor, covars = NULL, relative = TRUE, p.adj = "fdr", allResults = FALSE, ...){

diff --git a/R/DA.bay.R b/R/DA.bay.R
@@ -6,6 +6,15 @@
 #' @param allResults If TRUE will return raw results from the \code{getLikelihoods} function
 #' @param ... Additional arguments to the \code{getPriors.NB} and \code{getLikelihoods} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(100, size = 0.1, mu = 500), nrow = 50, ncol = 10)
+#' rownames(mat) <- 1:50
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running baySeq
+#' res <- DA.bay(data = mat, predictor = pred)
 #' @export
 
 DA.bay <- function(data, predictor, allResults = FALSE, ...){

diff --git a/R/DA.ds2.R b/R/DA.ds2.R
@@ -13,6 +13,15 @@
 #' @param allResults If TRUE will return raw results from the \code{DESeq} function
 #' @param ... Additional arguments for the \code{DESeq} function
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(200, size = 0.1, mu = 500), nrow = 20, ncol = 10)
+#' rownames(mat) <- 1:20
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running DESeq2
+#' res <- DA.ds2(data = mat, predictor = pred)
 #' @export
 
 DA.ds2 <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", coeff = 2, coeff.ref = 1, allResults = FALSE, ...){

diff --git a/R/DA.ds2x.R b/R/DA.ds2x.R
@@ -12,6 +12,15 @@
 #' @param allResults If TRUE will return raw results from the \code{DESeq} function
 #' @param ... Additional arguments for the \code{DESeq} function
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(6)
+#' mat <- matrix(rnbinom(200, size = 0.2, mu = 500), nrow = 20, ncol = 10)
+#' rownames(mat) <- 1:20
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running DESeq2
+#' res <- DA.ds2x(data = mat, predictor = pred)
 #' @export
 
 DA.ds2x <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", coeff = 2, coeff.ref = 1, allResults = FALSE, ...){

diff --git a/R/DA.ere.R b/R/DA.ere.R
@@ -6,6 +6,15 @@
 #' @param p.adj Character. P-value adjustment. Default "fdr". See \code{p.adjust} for details
 #' @param ... Additional arguments for the \code{calcNormFactors}, \code{estimateCommonDisp}, \code{estimateTagwiseDisp} and \code{exactTest} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running edgeR
+#' res <- DA.ere(data = mat, predictor = pred)
 #' @export
 
 DA.ere <- function(data, predictor, p.adj = "fdr", ...){

diff --git a/R/DA.ere2.R b/R/DA.ere2.R
@@ -6,6 +6,15 @@
 #' @param p.adj Character. P-value adjustment. Default "fdr". See \code{p.adjust} for details
 #' @param ... Additional arguments for the \code{calcNormFactors}, \code{estimateCommonDisp}, \code{estimateTagwiseDisp} and \code{exactTest} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(5)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running edgeR
+#' res <- DA.ere2(data = mat, predictor = pred)
 #' @export
 
 DA.ere2 <- function(data, predictor, p.adj = "fdr", ...){

diff --git a/R/DA.erq.R b/R/DA.erq.R
@@ -11,6 +11,15 @@
 #' @param allResults If TRUE will return raw results from the \code{glmQLFTest} function
 #' @param ... Additional arguments for the \code{calcNormFactors}, \code{estimateDisp}, \code{glmQLFit} and \code{glmQLFTest} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running edgeR
+#' res <- DA.erq(data = mat, predictor = pred)
 #' @export
 
 DA.erq <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", coeff = 2, allResults = FALSE, ...){

diff --git a/R/DA.erq2.R b/R/DA.erq2.R
@@ -11,6 +11,15 @@
 #' @param allResults If TRUE will return raw results from the \code{glmQLFTest} function
 #' @param ... Additional arguments for the \code{calcNormFactors}, \code{estimateDisp}, \code{glmQLFit} and \code{glmQLFTest} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(5)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running edgeR
+#' res <- DA.erq2(data = mat, predictor = pred)
 #' @export
 
 DA.erq2 <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", coeff = 2, allResults = FALSE, ...){

diff --git a/R/DA.fri.R b/R/DA.fri.R
@@ -9,6 +9,16 @@
 #' @param allResults If TRUE will return raw results from the \code{friedman.test} function
 #' @param ... Additional arguments for the \code{friedman.test} function
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table, predictor, and paired variable
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' subject <- rep(1:5, 3)
+#' 
+#' # Running Friedman test on each feature
+#' res <- DA.fri(data = mat, predictor = pred, paired = subject)
 #' @export
 
 DA.fri <- function(data, predictor, paired = NULL, relative = TRUE, p.adj = "fdr", allResults = FALSE, ...){
@@ -44,9 +54,9 @@ DA.fri <- function(data, predictor, paired = NULL, relative = TRUE, p.adj = "fdr
   if(allResults){
     return(reslist)
   } else {
-    res <- data.frame(statistic = vapply(reslist, function(x) x$statistic),
-                      parameter = vapply(reslist, function(x) x$parameter),
-                      pval = vapply(reslist, function(x) x$p.value))
+    res <- data.frame(statistic = sapply(reslist, function(x) x$statistic),
+                      parameter = sapply(reslist, function(x) x$parameter),
+                      pval = sapply(reslist, function(x) x$p.value))
     res$pval.adj <- p.adjust(res$pval, method = p.adj)
 
     res$Feature <- gsub(".Friedman.*","",rownames(res))

diff --git a/R/DA.kru.R b/R/DA.kru.R
@@ -8,6 +8,15 @@
 #' @param allResults If TRUE will return raw results from the \code{kruskal.test} function
 #' @param ... Additional arguments for the \code{kruskal.test} function
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running Kruskal-Wallis on each feature
+#' res <- DA.kru(data = mat, predictor = pred)
 #' @export
 
 DA.kru <- function(data, predictor, relative = TRUE, p.adj = "fdr", allResults = FALSE, ...){
@@ -41,7 +50,7 @@ DA.kru <- function(data, predictor, relative = TRUE, p.adj = "fdr", allResults =
   if(allResults){
     return(tests)
   } else {
-    res <- data.frame(pval = vapply(tests, function(x) x$p.value))
+    res <- data.frame(pval = sapply(tests, function(x) x$p.value))
     res$pval.adj <- p.adjust(res$pval, method = p.adj)
     res$Feature <- rownames(res)
     res$Method <- "Kruskal-Wallis (kru)" 

diff --git a/R/DA.lao.R b/R/DA.lao.R
@@ -10,6 +10,15 @@
 #' @param allResults If TRUE will return raw results from the \code{aov} function
 #' @param ... Additional arguments for the \code{aov} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running ANOVA on each feature
+#' res <- DA.lao(data = mat, predictor = pred)
 #' @export
 
 DA.lao <- function(data, predictor, covars = NULL, relative = TRUE, p.adj = "fdr", delta = 1, allResults = FALSE, ...){

diff --git a/R/DA.lao2.R b/R/DA.lao2.R
@@ -9,6 +9,15 @@
 #' @param allResults If TRUE will return raw results from the \code{aov} function
 #' @param ... Additional arguments for the \code{aov} functions
 #' @return A data.frame with with results.
+#' @examples
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1500, size = 0.1, mu = 500), nrow = 100, ncol = 15)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("A", 5), rep("B", 5), rep("C", 5))
+#' 
+#' # Running ANOVA on each feature
+#' res <- DA.lao2(data = mat, predictor = pred)
 #' @export
 
 DA.lao2 <- function(data, predictor, covars = NULL, p.adj = "fdr", delta = 0.001, allResults = FALSE, ...){

diff --git a/R/DA.lia.R b/R/DA.lia.R
@@ -12,6 +12,15 @@
 #' @param allResults If TRUE will return raw results from the \code{eBayes} function
 #' @param ... Additional arguments for the \code{eBayes} and \code{lmFit} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running limma
+#' res <- DA.lia(data = mat, predictor = pred)
 #' @export
 
 DA.lia <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", delta = 1, coeff = 2, allResults = FALSE,  ...){
@@ -46,7 +55,7 @@ DA.lia <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL
     }
 
     # Zero-correction
-    count_table <- apply(count_table, 2, function(y) vapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
+    count_table <- apply(count_table, 2, function(y) sapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
     if(any(count_table <= 0)) stop("Zero-correction failed. Dataset likely contains too many zeroes")
 
     # ALR transformation

diff --git a/R/DA.lic.R b/R/DA.lic.R
@@ -12,6 +12,15 @@
 #' @param allResults If TRUE will return raw results from the \code{eBayes} function
 #' @param ... Additional arguments for the \code{eBayes} and \code{lmFit} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running limma
+#' res <- DA.lic(data = mat, predictor = pred)
 #' @export
 
 DA.lic <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL, p.adj = "fdr", delta = 1, coeff = 2, allResults = FALSE,  ...){
@@ -46,7 +55,7 @@ DA.lic <- function(data, predictor, paired = NULL, covars = NULL, out.all = NULL
     }
 
     # Zero-correction
-    count_table <- apply(count_table, 2, function(y) vapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
+    count_table <- apply(count_table, 2, function(y) sapply(y,function(x) ifelse(x==0,delta,(1-(sum(y==0)*delta)/sum(y))*x)))
     if(any(count_table <= 0)) stop("Zero-correction failed. Dataset likely contains too many zeroes")
 
     # ALR transformation

diff --git a/R/DA.lim.R b/R/DA.lim.R
@@ -11,6 +11,15 @@
 #' @param allResults If TRUE will return raw results from the \code{eBayes} function
 #' @param ... Additional arguments for the \code{eBayes} and \code{lmFit} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running limma
+#' res <- DA.lim(data = mat, predictor = pred)
 #' @export
 
 DA.lim <- function(data, predictor, paired = NULL, covars = NULL, relative = TRUE, out.all = NULL, p.adj = "fdr", coeff = 2, allResults = FALSE, ...){

diff --git a/R/DA.lli.R b/R/DA.lli.R
@@ -13,6 +13,15 @@
 #' @param allResults If TRUE will return raw results from the \code{eBayes} function
 #' @param ... Additional arguments for the \code{eBayes} and \code{lmFit} functions
 #' @return A data.frame with with results.
+#' @examples 
+#' # Creating random count_table and predictor
+#' set.seed(4)
+#' mat <- matrix(rnbinom(1000, size = 0.1, mu = 500), nrow = 100, ncol = 10)
+#' rownames(mat) <- 1:100
+#' pred <- c(rep("Control", 5), rep("Treatment", 5))
+#' 
+#' # Running limma
+#' res <- DA.lli(data = mat, predictor = pred)
 #' @export
 
 DA.lli <- function(data, predictor, paired = NULL, covars = NULL, relative = TRUE, out.all = NULL, p.adj = "fdr", delta = 1, coeff = 2, allResults = FALSE,  ...){