Updated R version to 3.5

Updated NAMESPACE and DESCRIPTION to include SummarizedExperiment class
Updated R code, man, test, and vignette to include SummarizedExperiment class
Addressed minor formatting and warnings/notes

Author: Nathan Lawlor

DESCRIPTION

@@ -19,13 +19,14 @@ Description: Iteratively Adjusted Surrogate Variable Analysis (IA-SVA) is a
     iii), if significant, use the estimated factor as an additional known
     factor in the next iteration to uncover further hidden factors.
 Depends: 
-    R (>= 3.2),
+    R (>= 3.5),
 Imports:
     irlba,
     parallel,
     stats,
     cluster,
-    graphics
+    graphics,
+    SummarizedExperiment
 License: GPL-2
 biocViews: Preprocessing, QualityControl, BatchEffect
 Suggests:

NAMESPACE

@@ -4,6 +4,8 @@ export(fast_iasva)
 export(find_markers)
 export(iasva)
 export(study_R2)
+importFrom(SummarizedExperiment,SummarizedExperiment)
+importFrom(SummarizedExperiment,assay)
 importFrom(cluster,silhouette)
 importFrom(graphics,Axis)
 importFrom(graphics,mtext)

R/fast_iasva.R

@@ -1,17 +1,19 @@
 #' A function for fast IA-SVA 
 #'
 #' The iterative procedure of fast IA-SVA is implemented in this 
-#' function (fast_iasva). fast_iasva() iterativel identify a hidden factor
-#' for unwanted variation while accounting for all known factors, 
-#' compute its contribution  (i.e., the percentage of unmodeled variation
+#' function (fast_iasva). fast_iasva() iteratively identifies a hidden factor
+#' for unwanted variation while accounting for all known factors, and
+#' computes its contribution  (i.e., the percentage of unmodeled variation
 #' explained by the hidden factor) on the unmodeled variation in the data. 
 #' If the contribution is greater than a user-defined cutoff (pct.cutoff,
 #'  default = 1%), the factor is retained and used as a known variable in 
 #'  the next iteration to find further hidden factors.  
 #' @importFrom irlba irlba
+#' @importFrom SummarizedExperiment SummarizedExperiment assay
 #'
-#' @param Y read counts matrix with samples in row and genes in column.
-#' @param X  known variables. 
+#' @param Y A SummarizedExperiment class containing read counts where
+#' rows represent genes and columns represent samples.
+#' @param X  A design matrix of known variables (e.g., patient ID, gender).
 #' @param intercept If intercept = FALSE, the linear 
 #' intercept is not included in the model.
 #' @param num.sv number of surrogate variables to estimate. 
@@ -33,7 +35,6 @@
 #' @return n.sv number of obtained surrogate variables. 
 #' 
 #' @examples
-#' library(iasva)
 #' counts_file <- system.file("extdata", "iasva_counts_test.Rds",
 #'  package = "iasva")
 #' counts <- readRDS(counts_file)
@@ -43,11 +44,14 @@
 #' Geo_Lib_Size <- colSums(log(counts + 1))
 #' Patient_ID <- anns$Patient_ID
 #' mod <- model.matrix(~Patient_ID + Geo_Lib_Size)
-#' iasva.res <- fast_iasva(t(counts), mod[, -1], num.sv = 5)
+#' summ_exp <- SummarizedExperiment::SummarizedExperiment(assays = counts)
+#' iasva.res <- fast_iasva(summ_exp, mod[, -1], num.sv = 5)
 #' @export
 
 fast_iasva <- function(Y, X, intercept = TRUE, num.sv = NULL, pct.cutoff = 1,
                        num.tsv = NULL, tol = 1e-10, verbose = FALSE) {
+  # transpose the read counts
+  Y <- t(assay(Y))
   cat("fast IA-SVA running...")
   if (min(Y) < 0) {
     Y <- Y + abs(min(Y))
@@ -118,7 +122,7 @@ fast_iasva <- function(Y, X, intercept = TRUE, num.sv = NULL, pct.cutoff = 1,
     cat(paste0("\nSV", isv, " Detected!"))
   }
   if (isv > 0) {
-    colnames(sv) <- paste0("SV", 1:ncol(sv))
+    colnames(sv) <- paste0("SV", seq(from = 1, to = ncol(sv), by = 1))
     cat(paste0("\n# of obtained surrogate variables: ", length(pct)))
     return(list(sv = sv, pct = pct, n.sv = length(pct)))
   } else {

R/find_markers.R

@@ -5,8 +5,10 @@
 #'  identifies marker genes highly correlated with each surrogate variable
 #'  and returns a read counts matrix of the markers.
 #' @importFrom stats lm p.adjust
+#' @importFrom SummarizedExperiment SummarizedExperiment assay
 #'  
-#' @param Y read counts matrix with samples in row and genes in column.
+#' @param Y A SummarizedExperiment class containing read counts where
+#' rows represent genes and columns represent samples.
 #' @param iasva.sv  matrix of estimated surrogate variables,
 #'  one column for each surrogate variable. 
 #' @param method multiple testing adjustment method, default = "BH".
@@ -18,7 +20,6 @@
 #'  one column for each cell.
 #' 
 #' @examples
-#' library(iasva)
 #' counts_file <- system.file("extdata", "iasva_counts_test.Rds",
 #'  package = "iasva")
 #' counts <- readRDS(counts_file)
@@ -28,12 +29,15 @@
 #' Geo_Lib_Size <- colSums(log(counts + 1))
 #' Patient_ID <- anns$Patient_ID
 #' mod <- model.matrix(~Patient_ID + Geo_Lib_Size)
-#' iasva.res <- iasva(t(counts), mod[, -1], num.sv = 5, permute = FALSE)
-#' markers <- find_markers(t(counts), iasva.res$sv)
+#' summ_exp <- SummarizedExperiment::SummarizedExperiment(assays = counts)
+#' iasva.res <- iasva(summ_exp, mod[, -1], num.sv = 5, permute = FALSE)
+#' markers <- find_markers(summ_exp, iasva.res$sv)
 #' @export
 
 find_markers <- function(Y, iasva.sv, method = "BH", sig.cutoff = 0.05,
                          rsq.cutoff = 0.3, verbose = FALSE) {
+  # transpose the read counts
+  Y <- t(assay(Y))
   if (min(Y) < 0) {
     Y <- Y + abs(min(Y))
   }

R/iasva.R

@@ -7,9 +7,12 @@
 #'  unmodeled variation in the data. If the test statistic of detected factor
 #'  is significant, iasva() includes the factor as a known variable in the 
 #'  next iteration to find further hidden factors.  
-#'
-#' @param Y read counts matrix with samples in row and genes in column.
-#' @param X  known variables including the primary variables of interest. 
+#'  
+#' @importFrom SummarizedExperiment SummarizedExperiment assay
+#' @param Y A SummarizedExperiment class containing read counts where
+#' rows represent genes and columns represent samples.
+#' @param X  A model matrix of known variables
+#'  including the primary variables of interest. 
 #' @param intercept If intercept = FALSE, the linear intercept
 #'  is not included in the model.
 #' @param num.sv number of surrogate variables to estimate. 
@@ -36,7 +39,6 @@
 #' @return n.sv number of significant/obtained surrogate variables. 
 #' 
 #' @examples
-#' library(iasva)
 #' counts_file <- system.file("extdata", "iasva_counts_test.Rds",
 #'  package = "iasva")
 #' counts <- readRDS(counts_file)
@@ -46,7 +48,8 @@
 #' Geo_Lib_Size <- colSums(log(counts + 1))
 #' Patient_ID <- anns$Patient_ID
 #' mod <- model.matrix(~Patient_ID + Geo_Lib_Size)
-#' iasva.res<- iasva(t(counts), mod[, -1],verbose = FALSE, 
+#' summ_exp <- SummarizedExperiment::SummarizedExperiment(assays = counts)
+#' iasva.res<- iasva(summ_exp, mod[, -1],verbose = FALSE, 
 #'  permute = FALSE, num.sv = 5)
 #' @export
 
@@ -78,7 +81,7 @@ iasva <- function(Y, X, intercept = TRUE, num.sv = NULL, permute = TRUE,
     cat(paste0("\nSV", isv, " Detected!"))
   }
   if (isv > 0) {
-    colnames(sv) <- paste0("SV", 1:ncol(sv))
+    colnames(sv) <- paste0("SV", seq(from = 1, to = ncol(sv)))
     cat(paste0("\n# of significant surrogate variables: ", length(pval)))
     return(list(sv = sv, pc.stat.obs = pc.stat.obs,
                 pval = pval, n.sv = length(pval)))

R/iasva_unit.R

@@ -4,6 +4,8 @@
 iasva_unit <- function(Y, X, intercept = TRUE, permute = TRUE, num.p = 100,
                        threads = 1, num.sv.permtest = NULL,
                        tol = 1e-10, verbose = FALSE) {
+  # transpose the read counts
+  Y <- t(assay(Y))
   if (min(Y) < 0) {
     Y <- Y + abs(min(Y))
   }
@@ -60,13 +62,14 @@ iasva_unit <- function(Y, X, intercept = TRUE, permute = TRUE, num.p = 100,
     if (threads > 1) {
       threads <- min(threads, detectCores() - 1)
       cl <- makeCluster(threads)
-      pc.stat.null.vec <- tryCatch(parSapply(cl, 1:num.p, permute.svd),
+      pc.stat.null.vec <- tryCatch(parSapply(cl, 
+                                  seq(from = 1, to = num.p), permute.svd),
                                    error = function(err) {
                                      stopCluster(cl); stop(err)
                                      })
       stopCluster(cl)
     } else {
-      pc.stat.null.vec <- sapply(1:num.p, permute.svd)
+      pc.stat.null.vec <- vapply(seq(from = 1, to = num.p), permute.svd)
     }
     if (verbose) {
       cat("\n Empirical null distribution of the PC statistic:",

R/study_R2.R

@@ -8,8 +8,10 @@
 #' @importFrom graphics Axis mtext par plot
 #' @importFrom stats .lm.fit cutree dist hclust lm p.adjust resid
 #' @importFrom cluster silhouette
+#' @importFrom SummarizedExperiment SummarizedExperiment assay
 #'
-#' @param Y read counts matrix with samples in row and genes in column.
+#' @param Y A SummarizedExperiment class containing read counts where
+#' rows represent genes and columns represent samples.
 #' @param iasva.sv  matrix of estimated surrogate variables,
 #'  one column for each surrogate variable. 
 #' @param selected.svs list of SVs that are selected for the
@@ -21,7 +23,6 @@
 #'  as a function of R2 and corresponding matrices. 
 #'  
 #' @examples
-#' library(iasva)
 #' counts_file <- system.file("extdata", "iasva_counts_test.Rds",
 #'  package = "iasva")
 #' counts <- readRDS(counts_file)
@@ -31,10 +32,11 @@
 #' Geo_Lib_Size <- colSums(log(counts + 1))
 #' Patient_ID <- anns$Patient_ID
 #' mod <- model.matrix(~Patient_ID + Geo_Lib_Size)
-#' iasva.res<- iasva(t(counts), mod[, -1],verbose = FALSE, 
+#' summ_exp <- SummarizedExperiment::SummarizedExperiment(assays = counts)
+#' iasva.res<- iasva(summ_exp, mod[, -1],verbose = FALSE, 
 #' permute = FALSE, num.sv = 5)
 #' iasva.sv <- iasva.res$sv
-#' study_res <- study_R2(t(counts), iasva.sv)
+#' study_res <- study_R2(summ_exp, iasva.sv)
 #' 
 #' @export
 
@@ -68,11 +70,11 @@ study_R2 <- function(Y, iasva.sv, selected.svs = 2,
        pch = 18, col = "blue", type = "b", lty = 2, cex = 2)
   Axis(1, at = seq(1, length(Number.of.genes)), side = 1,
        labels = seq(0.1, end.point, 0.05), las = 2)
-  par(new = T)
-  plot(C.scores, xlab = "", ylab = "", axes = F, pch = 18, col = "red",
+  par(new = TRUE)
+  plot(C.scores, xlab = "", ylab = "", axes = FALSE, pch = 18, col = "red",
        type = "b", lty = 2, cex = 2)
   Axis(side = 4)
   mtext(side = 4, line = 2, "Average Silhouette Score", col = "red")
-  par(new = F)
+  par(new = FALSE)
   return(output.matrix)
 }