Merge pull request #40 from jcyang34/dev_master

bug fix for existing issues
msraredon · Sep 23, 2024 · 0bd2876 · 0bd2876
2 parents e131412 + 3dd08a6
commit 0bd2876
Show file tree

Hide file tree

Showing 16 changed files with 153 additions and 147 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -7,7 +7,7 @@ Description: Cell-cell Interactions at Single-Cell Resolution
 License: GPL-3
 Encoding: UTF-8
 LazyData: yes
-RoxygenNote: 7.3.1
+RoxygenNote: 7.3.2
 VignetteBuilder: knitr
 Depends:
     R (>= 4.0)

diff --git a/R/RunCellToCell.R b/R/RunCellToCell.R
@@ -5,8 +5,8 @@
 #' The default assay of this object is called "CellToCell" to distinguish it from normal Seurat objects.
 #' Meta.data slots by default contain "SendingType" "ReceivingType" and "VectorType" information.
 #' 
-#' @param sys.small A filtered Seurat object. The active identity will be used to define populations for connectomic sampling and crossings. 
-#' @param ground.truth Ground truth signaling mechanisms present in sys.small.
+#' @param filtered.obj A filtered Seurat object. The active identity will be used to define populations for connectomic sampling and crossings. 
+#' @param ground.truth Ground truth signaling mechanisms present in filtered.obj.
 #' @param assay The assay to run the SCC transformation on. Defaults to "RNA."
 #' @param meta.data.to.map A character vector of metadata names present in the original object which will be carried to the SCC objects
 #' @param output_format string. Choice of the output format. "seurat" will output a list of seurat objects, "raw" will output a list of lists with raw interaction matrix and compiled metadata
@@ -16,7 +16,7 @@
 #' @export
 
 
-RunCellToCell <- function(sys.small,
+RunCellToCell <- function(filtered.obj,
                           ground.truth,
                           assay,
                           meta.data.to.map,
@@ -25,13 +25,13 @@ RunCellToCell <- function(sys.small,
 
   ### CREATE MAPPING ###
 
-  # jc: Identify celltypes:names(table(Idents(sys.small))). Better to run check_celltypes, but harder to check
-  celltypes <- return_celltypes(sys.small)
+  # jc: Identify celltypes:names(table(Idents(filtered.obj))). Better to run check_celltypes, but harder to check
+  celltypes <- return_celltypes(filtered.obj)
 
   # Ligand dataset (listwise, for each celltype)
   lig.list <- list()
   for (i in 1:length(celltypes)){
-    temp <- subset(sys.small,idents = celltypes[i])
+    temp <- subset(filtered.obj,idents = celltypes[i])
     subunit.list <- list() # Builds sending (ligand) data for any number of ligand subunits
     for (s in 1:ncol(ground.truth$source.subunits)){ #For each subunit column...
       subunit.list[[s]] <- matrix(data = 1,nrow = nrow(ground.truth$source.subunits),ncol = ncol(temp)) #initialize a mechanism x barcode matrix of all NAs
@@ -47,7 +47,7 @@ RunCellToCell <- function(sys.small,
   # Receptor dataset (listwise, for each celltype)
   rec.list <- list()
   for (i in 1:length(celltypes)){
-    temp <- subset(sys.small,idents = celltypes[i])
+    temp <- subset(filtered.obj,idents = celltypes[i])
     subunit.list <- list() # Builds receiving (receptor) data for any number of receptor subunits
     for (t in 1:ncol(ground.truth$target.subunits)){
       subunit.list[[t]] <- matrix(data = 1,nrow = nrow(ground.truth$target.subunits),ncol = ncol(temp)) #initialize a mechanism x barcode matrix of all NAs
@@ -72,7 +72,7 @@ RunCellToCell <- function(sys.small,
   for (i in 1:length(celltypes)){
 
     # Define maximum number of comparisons for each pairing
-    num <- as.data.frame(table(Seurat::Idents(sys.small)))
+    num <- as.data.frame(table(Seurat::Idents(filtered.obj)))
     num$sender.freq <- ncol(lig.list[[i]])
     rownames(num) <- num$Var1
     num <- num[,-1]
@@ -98,8 +98,8 @@ RunCellToCell <- function(sys.small,
     rownames(scc.data[[i]]) <- paste(rownames(lig.data[[i]]),rownames(rec.data[[i]]),sep = '—')
     colnames(scc.data[[i]]) <- paste(colnames(lig.data[[i]]),colnames(rec.data[[i]]),sep = '—')
 
-    sending.cell.idents[[i]] <- as.character(Seurat::Idents(sys.small)[colnames(lig.data[[i]])])
-    receiving.cell.idents[[i]] <- as.character(Seurat::Idents(sys.small)[colnames(rec.data[[i]])])
+    sending.cell.idents[[i]] <- as.character(Seurat::Idents(filtered.obj)[colnames(lig.data[[i]])])
+    receiving.cell.idents[[i]] <- as.character(Seurat::Idents(filtered.obj)[colnames(rec.data[[i]])])
 
   }
 
@@ -109,8 +109,8 @@ RunCellToCell <- function(sys.small,
   #Use this matrix to create a Seurat object:
   demo <- Seurat::CreateSeuratObject(counts = as.matrix(scc),assay = 'CellToCell')
   # JC: Seurat V5 will not create data slot automatically, the following step is to manually add this slot
-  if(SeuratObject::Version(demo) >= 5){
-    demo <- NormalizeData(demo,assay = "CellToCell")  # Seura Object need to be >= 5.0.1
+  if(SeuratObject::Version(demo) >= "5.0.0"){
+    demo <- Seurat::NormalizeData(demo,assay = "CellToCell")  # Seura Object need to be >= 5.0.1
     demo@assays$CellToCell@layers$data <- demo@assays$CellToCell@layers$counts # Seura Object need to be>= 5.0.1
 
   }
@@ -133,9 +133,9 @@ RunCellToCell <- function(sys.small,
   # Identify sending and receiving barcodes
   sending.barcodes <- colnames(do.call(cbind,lig.data)) #This can be simplified if the above SCC construction is simplified
   receiving.barcodes <- colnames(do.call(cbind,rec.data)) #This can be simplified if the above SCC construction is simplified
-  # Pull and format sending and receiving metadata    jc: possible bug, change object to sys.small
-  sending.metadata <- as.matrix(sys.small@meta.data[,meta.data.to.map,drop=FALSE][sending.barcodes,])
-  receiving.metadata <- as.matrix(sys.small@meta.data[,meta.data.to.map,drop=FALSE][receiving.barcodes,])
+  # Pull and format sending and receiving metadata    jc: possible bug, change object to filtered.obj
+  sending.metadata <- as.matrix(filtered.obj@meta.data[,meta.data.to.map,drop=FALSE][sending.barcodes,])
+  receiving.metadata <- as.matrix(filtered.obj@meta.data[,meta.data.to.map,drop=FALSE][receiving.barcodes,])
   # Make joint metadata
   datArray <- abind::abind(sending.metadata,receiving.metadata,along=3)
   joint.metadata <- as.matrix(apply(datArray,1:2,function(x)paste(x[1],"-",x[2])))
@@ -154,7 +154,7 @@ RunCellToCell <- function(sys.small,
   Seurat::Idents(demo) <- demo$VectorType
 
   # How many vectors were captured by this sampling?
-  message(paste("\n",ncol(demo),'Cell-To-Cell edges computed, sampling',length(unique(demo$VectorType)),'distinct VectorTypes, out of',length(table(Seurat::Idents(sys.small)))^2,'total possible'))
+  message(paste("\n",ncol(demo),'Cell-To-Cell edges computed, sampling',length(unique(demo$VectorType)),'distinct VectorTypes, out of',length(table(Seurat::Idents(filtered.obj)))^2,'total possible'))
 
   if(output_format == "seurat") return(demo)
   else{

diff --git a/R/RunCellToCellSpatial.R b/R/RunCellToCellSpatial.R
@@ -1,7 +1,7 @@
 #' RunCellToCellSpatial
 #'
-#' @param sys.small A filtered Seurat object. The active identity will be used to define populations for connectomic sampling and crossings.
-#' @param ground.truth Ground truth signaling mechanisms present in sys.small.
+#' @param filtered.obj A filtered Seurat object. The active identity will be used to define populations for connectomic sampling and crossings.
+#' @param ground.truth Ground truth signaling mechanisms present in filtered.obj.
 #' @param assay The assay to run the SCC transformation on. Defaults to "RNA."
 #' @param meta.data.to.map A character vector of metadata names present in the original object which will be carried to the NICHES objects
 #' @param edgelist data.frame. Each row is an directional edge between two spatially connected cells
@@ -11,7 +11,7 @@
 #' @importFrom dplyr %>%
 #' @export
 
-RunCellToCellSpatial <- function(sys.small,
+RunCellToCellSpatial <- function(filtered.obj,
                                  ground.truth,
                                  assay,
                                  meta.data.to.map,
@@ -22,30 +22,30 @@ RunCellToCellSpatial <- function(sys.small,
 
     # Make ligand matrix
 
-    #lig.data <- sys.small@assays[[assay]]@data[ligands,edgelist$from]
+    #lig.data <- filtered.obj@assays[[assay]]@data[ligands,edgelist$from]
 
     subunit.list <- list() # Builds sending (ligand) data for any number of ligand subunits
     for (s in 1:ncol(ground.truth$source.subunits)){ #For each subunit column...
-      subunit.list[[s]] <- matrix(data = 1,nrow = nrow(ground.truth$source.subunits),ncol = ncol(getSeuratAssay(sys.small,assay,"data")[,edgelist$from])) #initialize a mechanism x barcode matrix of all NAs
-      colnames(subunit.list[[s]]) <- colnames(getSeuratAssay(sys.small,assay,"data")[,edgelist$from])
+      subunit.list[[s]] <- matrix(data = 1,nrow = nrow(ground.truth$source.subunits),ncol = ncol(getSeuratAssay(filtered.obj,assay,"data")[,edgelist$from])) #initialize a mechanism x barcode matrix of all NAs
+      colnames(subunit.list[[s]]) <- colnames(getSeuratAssay(filtered.obj,assay,"data")[,edgelist$from])
       rownames(subunit.list[[s]]) <- rownames(ground.truth$source.subunits)
       non.na.indices <- !is.na(ground.truth$source.subunits[,s]) #Identify rows in the s-th column of the ground truth which are not NA
-      subunit.list[[s]][non.na.indices,] <- as.matrix(getSeuratAssay(sys.small,assay,"data")[ground.truth$source.subunits[non.na.indices,s],edgelist$from])   #For every row in the initialized matrix corresponding to the indices of the ground.truth which are not NA, replace with the rows from the Seurat object corresponding to the genes in the ground.truth at those indices
+      subunit.list[[s]][non.na.indices,] <- as.matrix(getSeuratAssay(filtered.obj,assay,"data")[ground.truth$source.subunits[non.na.indices,s],edgelist$from])   #For every row in the initialized matrix corresponding to the indices of the ground.truth which are not NA, replace with the rows from the Seurat object corresponding to the genes in the ground.truth at those indices
     }
     lig.data <- Reduce('*',subunit.list)
     rm(subunit.list)
 
   # Make receptor matrix
 
-    #rec.data <- sys.small@assays[[assay]]@data[receptors,edgelist$to]
+    #rec.data <- filtered.obj@assays[[assay]]@data[receptors,edgelist$to]
 
     subunit.list <- list() # Builds receiving (receptor) data for any number of receptor subunits
     for (t in 1:ncol(ground.truth$target.subunits)){
-      subunit.list[[t]] <- matrix(data = 1,nrow = nrow(ground.truth$target.subunits),ncol = ncol(getSeuratAssay(sys.small,assay,"data")[,edgelist$to])) #initialize a mechanism x barcode matrix of all NAs
-      colnames(subunit.list[[t]]) <- colnames(getSeuratAssay(sys.small,assay,"data")[,edgelist$to])
+      subunit.list[[t]] <- matrix(data = 1,nrow = nrow(ground.truth$target.subunits),ncol = ncol(getSeuratAssay(filtered.obj,assay,"data")[,edgelist$to])) #initialize a mechanism x barcode matrix of all NAs
+      colnames(subunit.list[[t]]) <- colnames(getSeuratAssay(filtered.obj,assay,"data")[,edgelist$to])
       rownames(subunit.list[[t]]) <- rownames(ground.truth$target.subunits)
       non.na.indices <- !is.na(ground.truth$target.subunits[,t]) #Identify rows in the t-th column of the ground truth which are not NA
-      subunit.list[[t]][non.na.indices,] <- as.matrix(getSeuratAssay(sys.small,assay,"data")[ground.truth$target.subunits[non.na.indices,t],edgelist$to])   #For every row in the initialized matrix corresponding to the indices of the ground.truth which are not NA, replace with the rows from the Seurat object corresponding to the genes in the ground.truth at those indices
+      subunit.list[[t]][non.na.indices,] <- as.matrix(getSeuratAssay(filtered.obj,assay,"data")[ground.truth$target.subunits[non.na.indices,t],edgelist$to])   #For every row in the initialized matrix corresponding to the indices of the ground.truth which are not NA, replace with the rows from the Seurat object corresponding to the genes in the ground.truth at those indices
     }
     rec.data <- Reduce('*',subunit.list)
     rm(subunit.list)
@@ -54,16 +54,16 @@ RunCellToCellSpatial <- function(sys.small,
   scc <- lig.data*rec.data
   rownames(scc) <- paste(rownames(lig.data),rownames(rec.data),sep = '—')
   colnames(scc) <- paste(colnames(lig.data),colnames(rec.data),sep = '—')
-  sending.cell.idents <- as.character(Seurat::Idents(sys.small)[colnames(lig.data)])
-  receiving.cell.idents <- as.character(Seurat::Idents(sys.small)[colnames(rec.data)])
+  sending.cell.idents <- as.character(Seurat::Idents(filtered.obj)[colnames(lig.data)])
+  receiving.cell.idents <- as.character(Seurat::Idents(filtered.obj)[colnames(rec.data)])
   dim(scc)
 
   # Use this matrix to create a Seurat object:
   demo <- Seurat::CreateSeuratObject(counts = as.matrix(scc),assay = 'CellToCellSpatial')
 
   # JC: Seurat V5 will not create data slot automatically, the following step is to manually add this slot
-  if(SeuratObject::Version(demo) >= 5){
-    demo <- NormalizeData(demo,assay = "CellToCellSpatial")  # Seura Object need to be >= 5.0.1
+  if(SeuratObject::Version(demo) >= "5.0.0"){
+    demo <- Seurat::NormalizeData(demo,assay = "CellToCellSpatial")  # Seura Object need to be >= 5.0.1
     demo@assays$CellToCellSpatial@layers$data <- demo@assays$CellToCellSpatial@layers$counts # Seura Object need to be >= 5.0.1
 
   }
@@ -87,9 +87,9 @@ RunCellToCellSpatial <- function(sys.small,
     sending.barcodes <- colnames(lig.data)
     receiving.barcodes <- colnames(rec.data)
     # Pull and format sending and receiving metadata
-    # jc: possible bug, change object to sys.small
-    sending.metadata <- as.matrix(sys.small@meta.data[,meta.data.to.map,drop=FALSE][sending.barcodes,])
-    receiving.metadata <- as.matrix(sys.small@meta.data[,meta.data.to.map,drop=FALSE][receiving.barcodes,])
+    # jc: possible bug, change object to filtered.obj
+    sending.metadata <- as.matrix(filtered.obj@meta.data[,meta.data.to.map,drop=FALSE][sending.barcodes,])
+    receiving.metadata <- as.matrix(filtered.obj@meta.data[,meta.data.to.map,drop=FALSE][receiving.barcodes,])
     # Make joint metadata
     datArray <- abind::abind(sending.metadata,receiving.metadata,along=3)
     joint.metadata <- as.matrix(apply(datArray,1:2,function(x)paste(x[1],"-",x[2])))
@@ -107,7 +107,7 @@ RunCellToCellSpatial <- function(sys.small,
   Seurat::Idents(demo) <- demo$VectorType
 
   # How many vectors were captured by this sampling?
-  message(paste("\n",length(unique(demo$VectorType)),'distinct VectorTypes were computed, out of',length(table(Seurat::Idents(sys.small)))^2,'total possible'))
+  message(paste("\n",length(unique(demo$VectorType)),'distinct VectorTypes were computed, out of',length(table(Seurat::Idents(filtered.obj)))^2,'total possible'))
 
 
   if(output_format == "seurat") return(demo)