1.2.7

Author: Khodadadi-Jamayran

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: iCellR
 Type: Package
 Title: Analyzing High-Throughput Single Cell Sequencing Data
-Version: 1.2.5
+Version: 1.2.7
 Author: Alireza Khodadadi-Jamayran, 
   Joseph Pucella,
   Hua Zhou,
@@ -23,4 +23,4 @@ RoxygenNote: 6.1.1
 URL: https://github.com/rezakj/iCellR
 Suggests: phateR, Rmagic, Seurat
 NeedsCompilation: no
-Packaged: 2019-11-04 21:36:26 UTC; khodaa01
+Packaged: 2019-12-04 18:31:12 UTC; khodaa01

R/F008.norm.data.R

@@ -3,7 +3,7 @@
 #' This function takes an object of class iCellR and normalized the data based on "global.glsf", "ranked.glsf" or "spike.in" methods.
 #' @param x An object of class iCellR.
 #' @param norm.method Choose a normalization method, there are three option currently.
-#' Choose from "global.glsf", "ranked.glsf", "ranked.deseq", "deseq", "rpm","spike.in" or no.norm, default = "ranked.glsf".
+#' Choose from "global.glsf", "ranked.glsf","spike.in" or no.norm, default = "ranked.glsf".
 #' @param top.rank If the method is set to "ranked.glsf", you need to set top number of genes sorted based on global base mean, default = 500.
 #' @param spike.in.factors A numeric vector of spike-in values with the same cell id order as the main data.
 #' @param rpm.factor If the norm.method is set to "rpm" the library sizes would be diveded by this number, default = 1000 (higher numbers recomanded for bulk RNA-Seq).

R/F012.run.pca.R

@@ -7,7 +7,7 @@
 #' @param data.type Choose from "main" and "imputed", default = "main"
 #' @param plus.log.value A number to add to each value in the matrix before log transformasion to aviond Inf numbers, default = 0.1.
 #' @param gene.list A charactor vector of genes to be used for PCA. If "clust.method" is set to "gene.model", default = "my_model_genes.txt".
-#' @param batch.norm If TRUE the data will be normalized based on the genes in gene.list or top ranked genes.
+#' @param scale.data If TRUE the data will be scaled (log2 + plus.log.value), default = TRUE.
 #' @return An object of class iCellR.
 #' @examples
 #' demo.obj <- run.pca(demo.obj, method = "gene.model", gene.list = demo.obj@gene.model)
@@ -20,7 +20,7 @@ run.pca <- function (x = NULL,
                           method = "base.mean.rank",
                           top.rank = 500,
                           plus.log.value = 0.1,
-                          batch.norm = FALSE,
+                          scale.data = TRUE,
                           gene.list = "character") {
   if ("iCellR" != class(x)[1]) {
     stop("x should be an object of class iCellR")
@@ -36,8 +36,10 @@ run.pca <- function (x = NULL,
   # model base mean rank
   if (method == "base.mean.rank") {
     raw.data.order <- DATA[ order(rowMeans(DATA), decreasing = TRUE), ]
-    topGenes <- head(raw.data.order,top.rank)
-    TopNormLogScale <- log(topGenes + plus.log.value)
+    TopNormLogScale <- head(raw.data.order,top.rank)
+    if(scale.data == TRUE) {
+      TopNormLogScale <- log(TopNormLogScale + plus.log.value)
+    }
     # TopNormLogScale <- scale(topGenes)
 #    TopNormLogScale <- t(TopNormLogScale)
 #    TopNormLogScale <- as.data.frame(t(scale(TopNormLogScale)))
@@ -49,24 +51,28 @@ run.pca <- function (x = NULL,
     } else {
       genesForClustering <- gene.list
       topGenes <- subset(DATA, rownames(DATA) %in% genesForClustering)
-      if (batch.norm == FALSE){
         if (data.type == "main") {
-          TopNormLogScale <- log2(topGenes + plus.log.value)
-        }
+          TopNormLogScale <- topGenes
+          if(scale.data == TRUE) {
+            TopNormLogScale <- log(TopNormLogScale + plus.log.value)
+          }
         if (data.type == "imputed") {
-         # TopNormLogScale <- topGenes
-          TopNormLogScale <- t(scale(t(topGenes)))
+          TopNormLogScale <- topGenes
+          if(scale.data == TRUE) {
+            TopNormLogScale <- t(scale(t(topGenes)))
+#            TopNormLogScale <- log(TopNormLogScale + plus.log.value)
+          }
         }
       }
-      if (batch.norm == TRUE){
-        ## new method
-        libSiz <- colSums(topGenes)
-        norm.facts <- as.numeric(libSiz) / mean(as.numeric(libSiz))
-        dataMat <- as.matrix(topGenes)
-        normalized <- as.data.frame(sweep(dataMat, 2, norm.facts, `/`))
+#      if (batch.norm == TRUE){
+#        ## new method
+#        libSiz <- colSums(topGenes)
+#        norm.facts <- as.numeric(libSiz) / mean(as.numeric(libSiz))
+#        dataMat <- as.matrix(topGenes)
+#        normalized <- as.data.frame(sweep(dataMat, 2, norm.facts, `/`))
         # TopNormLogScale <- log2(normalized + plus.log.value)
-        TopNormLogScale <- normalized
-      }
+#        TopNormLogScale <- normalized
+#      }
     }
   }
 # Returns

R/F021.clust.avg.exp.R

@@ -2,13 +2,15 @@
 #'
 #' This function takes an object of class iCellR and creates an average gene expression for every cluster.
 #' @param x An object of class iCellR.
+#' @param data.type Choose from "main" and "imputed", default = "main"
 #' @return An object of class iCellR.
 #' @examples
 #' demo.obj <- clust.avg.exp(demo.obj)
 #'
 #' head(demo.obj@clust.avg)
 #' @export
-clust.avg.exp <- function (x = NULL) {
+clust.avg.exp <- function (x = NULL,
+                           data.type = "main") {
   if ("iCellR" != class(x)[1]) {
     stop("x should be an object of class iCellR")
   }
@@ -17,7 +19,14 @@ clust.avg.exp <- function (x = NULL) {
   sampleCondition <- DATA$clusters
   conditions <- sort(unique(sampleCondition))
   DATA1 <- DATA
-  Table = x@main.data
+  ## get main data
+  if (data.type == "main") {
+    Table <- x@main.data
+  }
+  if (data.type == "imputed") {
+    Table <- x@imputed.data
+  }
+#  Table = x@main.data
   for(i in conditions){
     IDs <- rownames(subset(DATA1, sampleCondition == i))
     DATA <- Table[ , which(names(Table) %in% IDs)]

R/F022.run.imput.R

@@ -3,24 +3,24 @@
 #' This function takes an object of class iCellR and runs imputation on the main data.
 #' @param x An object of class iCellR.
 #' @param imp.method Choose between "iCellR.imp" and "magic", defualt = "iCellR.imp".
-#' @param cell.ratio Percent of cells to use to find neighboring cells, default = 2.
+#' @param nn Number of neighboring cells to find, default = 10.
 #' @param dims PC dimentions to be used for the analysis, default = 10.
 #' @param data.type Choose between "tsne", "pca", "umap", "diffusion", default = "pca".
 #' @param genes character or integer vector, default: NULL vector of column names or column indices for which to return smoothed data If 'all_genes' or NULL, the entire smoothed matrix is returned
-#' @param k int, optional, default: 10 number of nearest neighbors on which to build kernel
-#' @param alpha int, optional, default: 15 sets decay rate of kernel tails. If NULL, alpha decaying kernel is not used
-#' @param t int, optional, default: 'auto' power to which the diffusion operator is powered sets the level of diffusion. If 'auto', t is selected according to the Procrustes disparity of the diffused data.'
+#' @param k if imp.method is magic; int, optional, default: 10 number of nearest neighbors on which to build kernel
+#' @param alpha if imp.method is magic; int, optional, default: 15 sets decay rate of kernel tails. If NULL, alpha decaying kernel is not used
+#' @param t if imp.method is magic; int, optional, default: 'auto' power to which the diffusion operator is powered sets the level of diffusion. If 'auto', t is selected according to the Procrustes disparity of the diffused data.'
 #' @param npca number of PCA components that should be used; default: 100.
 #' @param init magic object, optional object to use for initialization. Avoids recomputing intermediate steps if parameters are the same.
-#' @param t.max int, optional, default: 20 Maximum value of t to test for automatic t selection.
+#' @param t.max if imp.method is magic; int, optional, default: 20 Maximum value of t to test for automatic t selection.
 #' @param knn.dist.method string, optional, default: 'euclidean'. recommended values: 'euclidean', 'cosine' Any metric from 'scipy.spatial.distance' can be used distance metric for building kNN graph.
 #' @param verbose 'int' or 'boolean', optional (default : 1) If 'TRUE' or '> 0', message verbose updates.
 #' @param n.jobs 'int', optional (default: 1) The number of jobs to use for the computation. If -1 all CPUs are used. If 1 is given, no parallel computing code is used at all, which is useful for debugging. For n_jobs below -1, (n.cpus + 1 + n.jobs) are used. Thus for n_jobs = -2, all CPUs but one are used
 #' @param seed int or 'NULL', random state (default: 'NULL')
 #' @return An object of class iCellR.
 #' @export
 run.impute <- function (x = NULL,
-                        imp.method = "iCellR.imp", dims = 1:10,cell.ratio = 2,
+                        imp.method = "iCellR.imp", dims = 1:10, nn = 10,
                         data.type = "pca",genes = "all_genes", k = 10, alpha = 15, t = "auto",
                         npca = 100, init = NULL, t.max = 20,
                         knn.dist.method = "euclidean", verbose = 1, n.jobs = 1,
@@ -55,10 +55,11 @@ run.impute <- function (x = NULL,
     message(paste("   Calculating distance ..."))
     My.distances = as.matrix(dist(t(my.data.my.pca), method = knn.dist.method))
     ncells = dim(my.data)[2]
-    cell.num = ceiling(cell.ratio/100 * ncells)
-    message(paste("    ",cell.ratio,"percent of ",ncells, "cells is", cell.num))
-    message("     To change the number of neighboring cells cahnge cell.ratio option")
+#    cell.num = ceiling(cell.ratio/100 * ncells)
+    cell.num = nn
+#    message(paste("    ",cell.ratio,"percent of ",ncells, "cells is", cell.num))
     message(paste("    Finding",cell.num, "neighboring cells per cell ..."))
+    message("     To change the number of neighboring cells cahnge nn option")
     KNN1 = lapply(1:ncells, function(findKNN){
       order(My.distances[,findKNN])[1:cell.num]})
     ############

R/F027.clust.stats.plot.R

@@ -2,7 +2,7 @@
 #'
 #' This function takes an object of class iCellR and creates QC plot.
 #' @param x An object of class iCellR.
-#' @param plot.type Choose from "box.umi", "box.mito", "box.gene", default = "box.mito".
+#' @param plot.type Choose from "bar.cc", "pie.cc" , box.umi", "box.mito", "box.gene", default = "box.mito".
 #' @param cell.color Choose a color for points in the plot.
 #' @param cell.size A number for the size of the points in the plot, default = 1.
 #' @param box.color A color for the boxes in the "boxplot", default = "red".
@@ -41,7 +41,21 @@ clust.stats.plot <- function (x = NULL,
   MyClusts <- x@best.clust
   # merge
   DATA <- merge(DATA, MyClusts, by = "row.names", all.x=FALSE, all.y=TRUE)
-  # plot
+######  # plot
+# cell cycle
+  # bar
+  COUNTS <- c(1:length(DATA$Phase))
+  myBP <- ggplot(DATA,aes(y=COUNTS,
+                          x=clusters, fill = Phase)) +
+    geom_bar(stat = "identity") + theme_bw() +
+    theme(axis.text.x=element_text(angle=90)) +
+    ylab("Cell number ratio")
+  # pie
+  myPIE <- ggplot(DATA,aes(y=clusters, x="", fill = Phase)) +
+    geom_bar(stat = "identity", position = "fill") + theme_bw() + facet_wrap(~ clusters) +
+    theme(axis.title.y=element_blank(),
+          axis.text.y=element_blank(),
+          axis.ticks.y=element_blank()) + coord_polar(theta="y")
   # mito
   mito.percent.plot <- ggplot(DATA,aes(y=mito.percent, x=as.factor(clusters))) +
     geom_jitter(color = cell.color, size = cell.size, alpha = cell.transparency) +
@@ -70,7 +84,7 @@ clust.stats.plot <- function (x = NULL,
   if (plot.type == "box.umi") {
     if (interactive == TRUE) {
       OUT.PUT <- paste(out.name, ".html", sep="")
-      htmlwidgets::saveWidget(ggplotly(Mito.UMIs),OUT.PUT)
+      htmlwidgets::saveWidget(ggplotly(UMIsplot),OUT.PUT)
     }
     else
       return(UMIsplot)
@@ -79,7 +93,7 @@ clust.stats.plot <- function (x = NULL,
   if (plot.type == "box.mito") {
     if (interactive == TRUE) {
       OUT.PUT <- paste(out.name, ".html", sep="")
-      htmlwidgets::saveWidget(ggplotly(Mito.UMIs),OUT.PUT)
+      htmlwidgets::saveWidget(ggplotly(mito.percent.plot),OUT.PUT)
     }
     else
       return(mito.percent.plot)
@@ -88,9 +102,27 @@ clust.stats.plot <- function (x = NULL,
   if (plot.type == "box.gene") {
     if (interactive == TRUE) {
       OUT.PUT <- paste(out.name, ".html", sep="")
-      htmlwidgets::saveWidget(ggplotly(Mito.UMIs),OUT.PUT)
+      htmlwidgets::saveWidget(ggplotly(nGenes.plot),OUT.PUT)
     }
     else
       return(nGenes.plot)
   }
+  #
+  if (plot.type == "pie.cc") {
+    if (interactive == TRUE) {
+      OUT.PUT <- paste(out.name, ".html", sep="")
+      htmlwidgets::saveWidget(ggplotly(myPIE),OUT.PUT)
+    }
+    else
+      return(myPIE)
+  }
+  #
+  if (plot.type == "bar.cc") {
+    if (interactive == TRUE) {
+      OUT.PUT <- paste(out.name, ".html", sep="")
+      htmlwidgets::saveWidget(ggplotly(myBP),OUT.PUT)
+    }
+    else
+      return(myBP)
+  }
 }

R/F028.findMarkers.R

@@ -2,6 +2,7 @@
 #'
 #' This function takes an object of class iCellR and performs differential expression (DE) analysis to find marker genes for each cluster.
 #' @param x An object of class iCellR.
+#' @param data.type Choose from "main" and "imputed", default = "main"
 #' @param fold.change A number that designates the minimum fold change for out put, default = 2.
 #' @param padjval Minimum adjusted p value for out put, default = 0.1.
 #' @param Inf.FCs If set to FALSE the infinite fold changes would be filtered from out put, default = FALSE.
@@ -15,6 +16,7 @@
 #' head(marker.genes)
 #' @export
 findMarkers <- function (x = NULL,
+          data.type = "main",
           fold.change = 2,
           padjval = 0.1,
           Inf.FCs = FALSE,
@@ -25,8 +27,17 @@ findMarkers <- function (x = NULL,
   }
   ###########
 #  x <- clust.avg.exp(x)
-  dat <- x@main.data
+#  dat <- x@main.data
+  ## get main data
+  if (data.type == "main") {
+    dat <- x@main.data
+  }
+  if (data.type == "imputed") {
+    dat <- x@imputed.data
+  }
   # get cluster data
+  # get avrages
+  x <- clust.avg.exp(x, data.type = data.type)
   DATA <- x@best.clust
   ############## set wich clusters you want as condition 1 and 2
   MyClusts <- as.numeric(unique(DATA$clusters))
@@ -99,12 +110,14 @@ findMarkers <- function (x = NULL,
     mrgdall <- merge(Stats, Stats1, by="row.names")
     row.names(mrgdall) <- mrgdall$Row.names
     mrgdall <- mrgdall[,-1]
-    # get avrage data
-    AvData <- x@clust.avg
-    row.names(AvData) <- AvData$gene
-    mrgdall <- merge(mrgdall, AvData, by="row.names")
-    row.names(mrgdall) <- mrgdall$Row.names
-    mrgdall <- mrgdall[,-1]
+    ############################# get avrage data
+#    if (add.avg == TRUE) {
+      AvData <- x@clust.avg
+      row.names(AvData) <- AvData$gene
+      mrgdall <- merge(mrgdall, AvData, by="row.names")
+      row.names(mrgdall) <- mrgdall$Row.names
+      mrgdall <- mrgdall[,-1]
+#    }
     # make it an object
     DatNmaes=paste("DATAcluster",i,sep="_")
     eval(call("<-", as.name(DatNmaes), mrgdall))

R/F030.heatmap.gg.plot.R

@@ -38,7 +38,7 @@
 #' @export
 heatmap.gg.plot <- function (x = NULL,
                           gene = "NULL",
-                          cell.sort = TRUE,
+                          cell.sort = FALSE,
                           data.type = "main",
                           cluster.by = "clusters",
                           min.scale = -2.5,

R/F034.diff.exp.R

@@ -2,6 +2,7 @@
 #'
 #' This function takes an object of class iCellR and performs differential expression (DE) analysis for clusters and conditions.
 #' @param x An object of class iCellR.
+#' @param data.type Choose from "main" and "imputed", default = "main"
 #' @param de.by Choose from "clusters", "conditions", "clustBase.condComp" or "condBase.clustComp".
 #' @param cond.1 First condition to do DE analysis on.
 #' @param cond.2 Second condition to do DE analysis on.
@@ -14,6 +15,7 @@
 #'
 #' @export
 run.diff.exp <- function (x = NULL,
+                      data.type = "main",
                       de.by = "clusters",
                       cond.1 = "array",
                       cond.2 = "array",
@@ -22,7 +24,14 @@ run.diff.exp <- function (x = NULL,
     stop("x should be an object of class iCellR")
   }
   ###########
-  dat <- x@main.data
+#  dat <- x@main.data
+  ## get main data
+  if (data.type == "main") {
+    dat <- x@main.data
+  }
+  if (data.type == "imputed") {
+    dat <- x@imputed.data
+  }
   # 2 dimentions
   DATA <- x@best.clust
   ############## set wich clusters you want as condition 1 and 2

R/F045.cc.R

@@ -7,7 +7,9 @@
 #' @return The data frame object
 #' @importFrom Hmisc cut2
 #' @export
-cc <- function (object = NULL, s.genes = s.phase, g2m.genes = g2m.phase) {
+cc <- function (object = NULL,
+                s.genes = s.phase,
+                g2m.genes = g2m.phase) {
   if ("iCellR" != class(object)[1]) {
     stop("object should be an object of class iCellR")
   }