Develop

vizapp/functions.R

+LoadData <- function(dim=c("pca","tsne","umap"),dir,samp){
+  exp <- readMM(paste0(dir,samp,"/outs/filtered_feature_bc_matrix/matrix.mtx.gz"))
+  features <- read.table(paste0(dir,samp,"/outs/filtered_feature_bc_matrix/features.tsv.gz"),quote="\t")
+  barcodes <- read.table(paste0(dir,samp,"/outs/filtered_feature_bc_matrix/barcodes.tsv.gz"),quote="\t")
+  
+  exp.raw <- readMM(paste0(dir,samp,"/outs/raw_feature_bc_matrix/matrix.mtx.gz"))
+  features.raw <- read.table(paste0(dir,samp,"/outs/filtered_feature_bc_matrix/features.tsv.gz"),quote="\t")
+  barcodes.raw <- read.table(paste0(dir,samp,"/outs/raw_feature_bc_matrix/barcodes.tsv.gz"),quote="\t")
+  
+  dr <- list()
+  for (i in dim){
+    if (i=="pca"){
+      dr[[i]] <- read.csv(paste0(dir,samp,"/outs/analysis/",i,"/10_components/projection.csv"))
+      dr[[i]] <- dr[[i]][,1:3]
+    } else {
+      dr[[i]] <- read.csv(paste0(dir,samp,"/outs/analysis/",i,"/2_components/projection.csv"))
+    }
+  }
+  rm(i)
+  
+  cl <- c("graphclust",paste0("kmeans_",2:10,"_clusters"))
+  cluster <- list()
+  deg <- list()
+  for (i in cl){
+    cluster[[i]] <- read.csv(paste0(dir,samp,"/outs/analysis/clustering/",i,"/clusters.csv"))
+    deg[[i]] <- read.csv(paste0(dir,samp,"/outs/analysis/diffexp/",i,"/differential_expression.csv"))
+  }
+  rm(i)
+  
+  qc <- read.csv(paste0(dir,samp,"/outs/metrics_summary.csv"))
+  qc <- data.frame(t(qc))
+  qc[,2] <- qc[,1]
+  qc[,1] <- rownames(qc)
+  qc <- qc[-1,]
+  colnames(qc) <- c("Metric","Value")
+  
+  results <- list(
+    sample=samp,
+    exp=exp,
+    features=features,
+    barcodes=barcodes,
+    exp.raw=exp.raw,
+    features.raw=features.raw,
+    barcodes.raw=barcodes.raw,
+    dr=dr,
+    cluster=cluster,
+    deg=deg,
+    qc=qc
+  )
+  return(results)
+}
+
+Plot.cluster <- function(dr,cluster){
+  pl <-merge(dr,cluster,by="Barcode",all.x=TRUE)
+  axis.labs <- colnames(pl)[2:3]
+  colnames(pl)[2:3] <- c("dim1","dim2")
+  plot.out <- ggplot(pl,aes(x=dim1,y=dim2,col=factor(Cluster)))+geom_point()+scale_color_viridis(discrete=TRUE)+xlab(axis.labs[1])+ylab(axis.labs[2])+labs(col="Cluster")+theme_cowplot()
+  return(plot.out)
+}
+
+Plot.feature <- function(dr,ft){
+  dr$exp <- ft
+  axis.labs <- colnames(dr)[2:3]
+  colnames(dr)[2:3] <- c("dim1","dim2")
+  if (sum(dr$exp) != 0){
+    plot.out <- ggplot(dr,aes(x=dim1,y=dim2,col=exp))+geom_point()+scale_color_viridis(option="inferno")+xlab(axis.labs[1])+ylab(axis.labs[2])+labs(col="Expression")+theme_cowplot()
+  } else {
+    plot.out <- ggplot(dr,aes(x=dim1,y=dim2,col=exp))+geom_point()+scale_color_gradient(low="black",high="black")+xlab(axis.labs[1])+ylab(axis.labs[2])+labs(col="Expression")+theme_cowplot()
+  }
+  return(plot.out)
+}
+
+Plot.violinbox <- function(ft,cluster){
+  cluster$exp <- ft
+  cluster$Cluster <- factor(cluster$Cluster)
+  plot.out <- ggplot(cluster,aes(x=Cluster,y=exp,fill=Cluster))+geom_violin(scale="width",trim=TRUE)+geom_boxplot(width=0.25,fill="white",outlier.shape=NA)+scale_fill_viridis(discrete=TRUE)+ylab("Expression")+theme_cowplot()+theme(axis.text.x=element_text(angle=45))
+
+  return(plot.out)
+}
+
+Plot.cliffknee <- function(barcodes.raw,exp.raw,barcodes){
+  umi <- matrix(nrow=length(barcodes.raw[,1]),ncol=4)
+  umi[,1] <- t(colSums(exp.raw))
+  umi[,2] <- as.character(barcodes.raw[,1])
+  umi <- as.data.frame(umi)
+  umi[,1] <- as.numeric(levels(umi[,1]))[umi[,1]]
+  colnames(umi) <- c("nUMI","barcodes","rank","Cell")
+  umi <- umi[order(umi$nUMI,decreasing=TRUE),]
+  umi$rank <- 1:nrow(umi)
+  umi$Cell <- factor(as.character(umi$barcodes %in% barcodes[,1]))
+  umi$Cell <- factor(umi$Cell,levels(umi$Cell)[c(2,1)])
+  plot.out <- ggplot(umi,aes(x=rank,y=nUMI,col=Cell))+geom_point()+scale_color_manual(values=c("darkgreen","darkred"))+
+    scale_x_log10(breaks=trans_breaks("log10",function(x) 10^x),labels=trans_format("log10",math_format(10^.x)))+
+    scale_y_log10(breaks=trans_breaks("log10",function(x) 10^x),labels=trans_format("log10",math_format(10^.x)))+
+    annotation_logticks()+xlab("Barcode Rank")+ylab("nUMI")+theme_cowplot()
+  
+  return(plot.out)
+}