Merge pull request #195 from SpatialHackathon/data_Locus_coeruleus

Adding data for issue #38

Author: Jieran-S

data/locus_coeruleus/locus_coeruleus.R

@@ -0,0 +1,130 @@
+#!/usr/bin/env Rscript
+
+# Author_and_contribution: Niklas Mueller-Boetticher; created template
+# Author_and_contribution: Florian Heyl (@heylf); created code
+
+# H_E.json and H_E.tiff not public. Request for access is still unanswered.
+
+suppressPackageStartupMessages(library(optparse))
+suppressPackageStartupMessages(library(SpatialExperiment))
+suppressPackageStartupMessages(library(SingleCellExperiment))
+suppressPackageStartupMessages(library(WeberDivechaLCdata))
+suppressPackageStartupMessages(library(Matrix))
+
+option_list <- list(
+  make_option(
+    c("-o", "--out_dir"),
+    type = "character", default = NULL,
+    help = "Output directory to write files to."
+  )
+)
+
+description <- "Load data (Visium) for Locus_coeruleus from Lukas M. Weber at al. (2022); 
+The gene expression landscape of the human locus coeruleus revealed 
+by single-nucleus and spatially-resolved transcriptomics"
+
+opt_parser <- OptionParser(
+  usage = description,
+  option_list = option_list
+)
+opt <- parse_args(opt_parser)
+
+out_dir <- opt$out_dir
+
+args = commandArgs(trailingOnly=TRUE)
+
+# Load data
+spe <- WeberDivechaLCdata_Visium()
+
+coords <- as.data.frame(spe@int_colData$spatialCoords)
+colnames(coords) <- c('x', 'y')
+coords_rownames <- rownames(spe@int_colData$spatialCoords)
+
+counts <- spe@assays@data$counts
+
+counts_func <- function(x){
+  fields <- unlist(strsplit(x, "_"))
+  if ( length(fields) == 4 ){
+    return(paste(fields[1:3], collapse = "_"))
+  } else {
+    return(paste(fields[1:4], collapse = "_"))
+  }
+}
+counts_lc <- unlist(lapply(colnames(counts), counts_func))
+
+LC_samples <- unique(spe@colData$sample_id)
+
+for ( dir in LC_samples ){
+    dir <- paste0(out_dir, "/", dir)
+    if ( dir.exists(dir) == FALSE ){
+        dir.create(dir, recursive=TRUE)
+    }
+}
+
+print("Create output ...")
+
+patient_list <- c()
+sample_list <- c()
+directory_list <- c()
+
+# Write coordinates.tsv, observations.tsv, features.tsv, counts.mtx and labels.tsv
+for (lc in LC_samples){
+    
+    print(lc)
+    dir <- paste0(out_dir, "/", lc)
+
+    # Write coordinates.tsv
+    coords_subset <- coords[which(spe@colData$sample_id == lc),]
+    rownames(coords_subset) <- coords_rownames[which(spe@colData$sample_id == lc)]
+    write.table(coords_subset, file = paste0(dir, "/coordinates.tsv"), col.names = NA,
+                sep = "\t", quote = FALSE, row.names = TRUE)
+  
+    # Count matrix has rows = genes/features and cols = cells/observations
+    counts_subset <- counts[,which(counts_lc == lc)]
+    
+    # Transpose to have rows = cells/observations
+    counts_subset <- t(counts_subset)
+    writeMM(counts_subset, file = paste0(dir, "/counts.mtx"))
+  
+    observations_subset <- spe@colData[which(counts_lc == lc),]
+    rownames(observations_subset) <- lapply(rownames(observations_subset), function(x){tail(unlist(strsplit(x,"_")),1)})
+    write.table(observations_subset, file = paste0(dir, "/observations.tsv"), col.names = NA, sep = "\t", quote = FALSE)
+
+    labels <- spe@colData$annot_region[which(spe@colData$sample_id == lc)]
+    labels[which(labels == TRUE)] <- "LC"
+    labels[which(labels == FALSE)] <- "non_LC"
+
+    labels_df <- data.frame(label = labels)
+    rownames(labels_df) <- rownames(observations_subset)
+    write.table(labels_df, file = paste0(dir, "/labels.tsv"), col.names = NA, sep = "\t", quote = FALSE)
+
+    # Fill metadata
+    patient_list <- c(patient_list, as.character(unique(spe@colData[which(spe@colData$sample_id == lc),]$donor_id)))
+    sample_list <- c(sample_list, lc)
+    directory_list <- c(directory_list, dir)
+
+    # Write features.tsv
+    features <- as.data.frame(spe@rowRanges)
+    rownames(features) <- spe@rowRanges$gene_id
+    write.table(features, file = paste0(dir,"/features.tsv"), col.names = NA, sep = "\t", quote = FALSE)
+}
+
+## Metadata files
+samples_df <- data.frame(
+  patient = patient_list, 
+  sample = sample_list, 
+  position = rep(NA, length(patient_list)), # Not sure what position means
+  replicate = rep(NA, length(patient_list)), # If they have replicated then it is really badly named
+  directory = directory_list, 
+  n_clusters = rep(2, length(patient_list)),
+  stringsAsFactors = FALSE
+)
+row.names(samples_df) <- NULL
+write.table(samples_df, file = file.path(out_dir, "samples.tsv"), sep = "\t", col.names = NA, quote = FALSE)
+
+technology = "Visium"
+json <- file(file.path(out_dir, "experiment.json"))
+writeLines(c(paste0('{"technology": "', technology, '"}')), json)
+close(json)
+
+print("...finished")

data/locus_coeruleus/locus_coeruleus.yml

@@ -0,0 +1,8 @@
+channels:
+  - conda-forge
+  - bioconda
+dependencies:
+  - r-base=4.3.1
+  - bioconductor-weberdivechalcdata=1.4.0
+  - r-optparse=1.7.3
+  - r-matrix=1.6.1