biopipen 0.21.0__py3-none-any.whl → 0.34.26__py3-none-any.whl

biopipen/scripts/scrna/CellTypeAnnotation-direct.R

@@ -1,48 +1,85 @@
-source("{{biopipen_dir}}/utils/misc.R")
 library(Seurat)
+library(rlang)
+library(dplyr)
+library(tidyseurat)
 
-sobjfile = {{in.sobjfile | r}}
-outfile = {{out.outfile | r}}
-celltypes = {{envs.cell_types | r}}
-newcol = {{envs.newcol | r}}
+sobjfile <- {{in.sobjfile | r}}
+outfile <- {{out.outfile | r}}
+celltypes <- {{envs.cell_types | r}}
+newcol <- {{envs.newcol | r}}
+ident <- {{envs.ident | r }}
+merge_same_labels <- {{envs.merge | r}}
+more_cell_types <- {{envs.more_cell_types | r}}
+
+log <- biopipen.utils::get_logger()
 
 if (is.null(celltypes) || length(celltypes) == 0) {
-    warning("No cell types are given!")
+    log$warn("No cell types are given!")
+    if (!is.null(more_cell_types) && length(more_cell_types) > 0) {
+        log$warn("`envs.celltypes` is not given, won't process `envs.more_cell_types`!")
+    }
 
+    if (merge_same_labels) {
+        log$warn("Ignoring 'envs.merge' because no cell types are given!")
+    }
     # create a symbolic link to the input file
     file.symlink(sobjfile, outfile)
 } else {
-    sobj = readRDS(sobjfile)
-    idents = as.character(unique(Idents(sobj)))
-    idents = idents[order(as.numeric(idents))]
-
-    if (length(celltypes) < length(idents)) {
-        celltypes = c(celltypes, idents[(length(celltypes) + 1):length(idents)])
-    } else if (length(celltypes) > length(idents)) {
-        celltypes = celltypes[1:length(idents)]
-        warning(
-            "The length of cell types is longer than the number of clusters!",
-            immediate. = TRUE
-        )
+    log$info("Loading Seurat object ...")
+    sobj <- biopipen.utils::read_obj(sobjfile)
+    ident <- ident %||% biopipen.utils::GetIdentityColumn(sobj)
+    Idents(sobj) <- ident
+    idents <- Idents(sobj)
+    if (is.factor(idents)) {
+        idents <- levels(idents)
+    } else {
+        idents <- as.character(unique(idents))
     }
-    for (i in seq_along(celltypes)) {
-        if (celltypes[i] == "-" || celltypes[i] == "") {
-            celltypes[i] = idents[i]
+    process_celltypes <- function(ct, key = NULL) {
+        if (length(ct) < length(idents)) {
+            ct <- c(ct, idents[(length(ct) + 1):length(idents)])
+        } else if (length(ct) > length(idents)) {
+            ct <- ct[1:length(idents)]
+            if (is.null(key)) {
+                log$warn("The length of cell types is longer than the number of clusters!")
+            } else {
+                log$warn(paste0("The length of cell types for '", key, "' is longer than the number of clusters!"))
+            }
+        }
+        for (i in seq_along(ct)) {
+            if (ct[i] == "-" || ct[i] == "") {
+                ct[i] <- idents[i]
+            }
         }
+        names(ct) <- idents
+        return(ct)
     }
-    names(celltypes) = idents
 
+    if (!is.null(more_cell_types) && length(more_cell_types) > 0) {
+        for (key in names(more_cell_types)) {
+            ct <- more_cell_types[[key]]
+            ct <- process_celltypes(ct, key)
+            log$info(paste0("Adding additional cell type annotation: '", key, "' ..."))
+            sobj@meta.data[[key]] <- ct[as.character(Idents(sobj))]
+        }
+    }
+
+    celltypes <- process_celltypes(celltypes)
+
+    log$info("Renaming cell types ...")
     if (is.null(newcol)) {
-        sobj$seurat_clusters_id = Idents(sobj)
-        celltypes$object = sobj
-        sobj = do_call(RenameIdents, celltypes)
-        sobj$seurat_clusters = Idents(sobj)
+        sobj <- rename_idents(sobj, ident, celltypes)
+        log$info("Filtering clusters if NA ...")
+        sobj <- filter(sobj, !!sym(ident) != "NA" & !is.na(!!sym(ident)))
     } else {
-        celltypes$object = sobj
-        sobj = do_call(RenameIdents, celltypes)
-        sobj[[newcol]] = Idents(sobj)
-        Idents(sobj) = "seurat_clusters"
+        sobj[[newcol]] <- celltypes[as.character(Idents(sobj))]
+    }
+
+    if (merge_same_labels) {
+        log$info("Merging clusters with the same labels ...")
+        sobj <- merge_clusters_with_same_labels(sobj, newcol)
     }
 
-    saveRDS(sobj, outfile)
+    log$info("Saving Seurat object ...")
+    biopipen.utils::save_obj(sobj, outfile)
 }

biopipen/scripts/scrna/CellTypeAnnotation-hitype.R

@@ -1,22 +1,26 @@
-library(Seurat)
+library(rlang)
 library(dplyr)
 library(hitype)
 
-source("{{biopipen_dir}}/utils/misc.R")
-
 sobjfile = {{in.sobjfile | r}}
 outfile = {{out.outfile | r}}
 tissue = {{envs.hitype_tissue | r}}
 db = {{envs.hitype_db | r}}
 newcol = {{envs.newcol | r}}
+ident = {{envs.ident | r }}
+merge_same_labels = {{envs.merge | r}}
 
 if (is.null(db)) { stop("`envs.hitype_db` is not set") }
 
-print("- Reading Seurat object...")
-sobj = readRDS(sobjfile)
+log <- get_logger()
+
+log$info("Reading Seurat object...")
+sobj = biopipen.utils::read_obj(sobjfile)
+ident <- ident %||% biopipen.utils::GetIdentityColumn(sobj)
+Idents(sobj) <- ident
 
 # prepare gene sets
-print("- Preparing gene sets...")
+log$info("Preparing gene sets...")
 if (startsWith(db, "hitypedb_") && !grepl(".", db, fixed = TRUE)) {
     gs_list = gs_prepare(eval(as.symbol(db)), tissue)
 } else {
@@ -24,33 +28,34 @@ if (startsWith(db, "hitypedb_") && !grepl(".", db, fixed = TRUE)) {
 }
 
 # run RunHitype
-print("- Running RunHitype...")
+log$info("Running RunHitype...")
 sobj = RunHitype(sobj, gs_list, threshold = 0.0, make_unique = TRUE)
 
-print("- Renaming cell types...")
-hitype_levels = sobj@meta.data %>%
-    select(seurat_clusters, hitype) %>%
-    distinct(seurat_clusters, .keep_all = TRUE) %>%
-    arrange(as.numeric(seurat_clusters)) %>%
-    pull("hitype")
+log$info("Renaming cell types...")
+hitype_labels <- sobj@meta.data %>%
+    distinct(!!sym(ident), hitype)
+hitype_labels <- split(hitype_labels$hitype, hitype_labels[[ident]])
 
 if (is.null(newcol)) {
-    sobj$seurat_clusters_id = sobj$seurat_clusters
-    sobj$seurat_clusters = factor(sobj$hitype, levels = hitype_levels)
-    Idents(sobj) = "seurat_clusters"
+    sobj <- rename_idents(sobj, ident, hitype_labels)
 } else {
-    sobj[[newcol]] = factor(sobj$hitype, levels = hitype_levels)
+    sobj[[newcol]] = sobj$hitype
+}
+
+if (merge_same_labels) {
+    log$info("Merging clusters with the same labels...")
+    sobj = merge_clusters_with_same_labels(sobj, newcol)
 }
 
-print("- Saving Seurat object...")
-saveRDS(sobj, outfile)
+log$info("Saving Seurat object...")
+biopipen.utils::save_obj(sobj, outfile)
 
-print("- Saving the mappings ...")
+log$info("Saving the mappings ...")
 if (is.null(newcol)) {
     celltypes = sobj@meta.data %>%
-        group_by(seurat_clusters_id) %>%
+        group_by(!!sym(backup_col)) %>%
         summarize(CellType = hitype[1]) %>%
-        select(Cluster = seurat_clusters_id, CellType) %>%
+        select(Cluster = !!sym(backup_col), CellType) %>%
         ungroup()
 } else {
     celltypes = sobj@meta.data %>%

biopipen/scripts/scrna/CellTypeAnnotation-sccatch.R

@@ -1,35 +1,42 @@
-source("{{biopipen_dir}}/utils/misc.R")
 library(scCATCH)
 library(Seurat)
+library(biopipen.utils)
 
 sobjfile = {{in.sobjfile | r}}
 outfile = {{out.outfile | r}}
 sccatch_args = {{envs.sccatch_args | r}}
 newcol = {{envs.newcol | r}}
+ident = {{envs.ident | r }}
+merge_same_labels = {{envs.merge | r}}
+
+log <- get_logger()
 
-if (is.null(sccatch_args$tissue)) { stop("`envs.sccatch_args.tissue` origin of cells must be defined.") }
-if (is.null(sccatch_args$species)) {
-    sccatch_args$species = "Human"
-}
 if (!is.null(sccatch_args$marker)) {
-    cellmatch = readRDS(sccatch_args$marker)
+    cellmatch = read_obj(sccatch_args$marker)
     sccatch_args$if_use_custom_marker = TRUE
 }
 sccatch_args$marker = cellmatch
 
-if (is.null(sccatch_args$cancer)) {
-    sccatch_args$cancer = "Normal"
-}
 if (is.integer(sccatch_args$use_method)) {
     sccatch_args$use_method = as.character(sccatch_args$use_method)
 }
 
-sobj = readRDS(sobjfile)
+log$info("Reading Seurat object...")
+sobj = read_obj(sobjfile)
+ident <- ident %||% GetIdentityColumn(sobj)
+Idents(sobj) <- ident
 
+log$info("Running createscCATCH ...")
 obj = createscCATCH(data = GetAssayData(sobj), cluster = as.character(Idents(sobj)))
 sccatch_args$object = obj
 
+log$info("Running findmarkergene ...")
 obj = do_call(findmarkergene, sccatch_args)
+
+log$info("Running findcelltype ...")
+obj = findcelltype(object = obj)
+
+log$info("Saving the mappings ...")
 write.table(
     obj@celltype,
     file = file.path(dirname(outfile), "cluster2celltype.tsv"),
@@ -41,18 +48,19 @@ celltypes = as.list(obj@celltype$cell_type)
 names(celltypes) = obj@celltype$cluster
 
 if (length(celltypes) == 0) {
-    warning("No cell types annotated from the database!")
+    log$warn("- No cell types annotated from the database!")
 } else {
     if (is.null(newcol)) {
-        sobj$seurat_clusters_id = Idents(sobj)
-        celltypes$object = sobj
-        sobj = do_call(RenameIdents, celltypes)
-        sobj$seurat_clusters = Idents(sobj)
+        sobj <- rename_idents(sobj, ident, celltypes)
     } else {
-        celltypes$object = sobj
-        sobj = do_call(RenameIdents, celltypes)
-        sobj[[newcol]] = Idents(sobj)
-        Idents(sobj) = "seurat_clusters"
+        sobj@meta.data[[newcol]] = celltypes[as.character(Idents(sobj))]
+    }
+
+    if (merge_same_labels) {
+        log$info("Merging clusters with the same labels ...")
+        sobj = merge_clusters_with_same_labels(sobj, newcol)
     }
 }
-saveRDS(sobj, outfile)
+
+log$info("Saving Seurat object ...")
+save_obj(sobj, outfile)

biopipen/scripts/scrna/CellTypeAnnotation-sctype.R

@@ -1,34 +1,43 @@
 library(dplyr)
 library(HGNChelper)
 library(Seurat)
+library(rlang)
+library(biopipen.utils)
 
-source("{{biopipen_dir}}/utils/misc.R")
-source("{{biopipen_dir}}/scripts/scrna/sctype.R")
+{% include biopipen_dir + "/scripts/scrna/sctype.R" %}
 
 sobjfile = {{in.sobjfile | r}}
 outfile = {{out.outfile | r}}
 tissue = {{envs.sctype_tissue | r}}
 db = {{envs.sctype_db | r}}
 newcol = {{envs.newcol | r}}
+ident = {{envs.ident | r }}
+merge_same_labels = {{envs.merge | r}}
 
 if (is.null(db)) { stop("`envs.sctype_args.db` is not set") }
 
-print("- Reading Seurat object...")
-sobj = readRDS(sobjfile)
+log <- get_logger()
+
+log$info("Reading Seurat object...")
+sobj = biopipen.utils::read_obj(sobjfile)
+ident <- ident %||% biopipen.utils::GetIdentityColumn(sobj)
+Idents(sobj) <- ident
 
 # prepare gene sets
-print("- Preparing gene sets...")
+log$info("Preparing gene sets...")
 gs_list = gene_sets_prepare(db, tissue)
 
-scRNAseqData = GetAssayData(sobj, slot = "scale.data")
+scRNAseqData = GetAssayData(sobj, layer = "scale.data")
 idents = as.character(unique(Idents(sobj)))
 idents = idents[order(as.numeric(idents))]
 
+log$info("Working on different levels of cell type labels ...")
 cell_types_list = list()
 for (i in seq_along(gs_list)) {
+    log$info("- Working on level {i} ...")
     if (is.null(gs_list[[i]])) next
 
-    print(paste0("- Calculating cell-type scores for level ", i, "..."))
+    log$info("  Calculating cell-type scores ...")
     es.max = sctype_score(
         scRNAseqData = scRNAseqData,
         scaled = TRUE,
@@ -36,7 +45,7 @@ for (i in seq_along(gs_list)) {
         gs2 = gs_list[[i]]$gs_negative
     )
 
-    print(paste0("- Merging cell-type scores by cluster for level ", i, "..."))
+    log$info("  Merging cell-type scores by cluster ...")
     cl_resutls = do_call(
         "rbind",
         lapply(
@@ -59,12 +68,12 @@ for (i in seq_along(gs_list)) {
         write("\n####### sctype_scores_count ########", stderr())
         write(capture.output(sctype_scores_count), stderr())
         write("\n####################################", stderr())
-        warning("Scores tied in the above clusters.", immediate. = TRUE)
+        log$info("  Scores tied in the above clusters.", immediate. = TRUE)
     }
 
     if (length(gs_list) == 1 || i > 1) {
         # set low-confident (low ScType score) clusters to "unknown"
-        print("- Setting low-confident clusters to 'Unknown'...")
+        log$info("  Setting low-confident clusters to 'Unknown'...")
         sctype_scores$type[as.numeric(as.character(sctype_scores$scores)) < sctype_scores$ncells/4] = "Unknown"
     }
 
@@ -82,7 +91,7 @@ for (i in seq_along(gs_list)) {
 if (length(cell_types_list) == 1) {
     celltypes = cell_types_list[[1]]
 } else {
-    print("- Merging cell types at all levels ...")
+    log$info("Merging cell types at all levels ...")
     celltypes = list()
 
     for (i in idents) {
@@ -97,24 +106,35 @@ if (length(cell_types_list) == 1) {
 }
 
 
-print("- Renaming cell types...")
+log$info("Renaming cell types...")
+ct_numbering = list()
+for (key in names(celltypes)) {
+    ct = celltypes[[key]]
+    ct_numbering[[ct]] = ct_numbering[[ct]] %||% 0
+    if (ct_numbering[[ct]] > 0) {
+        celltypes[[key]] = paste0(ct, ".", ct_numbering[[ct]])
+    }
+    ct_numbering[[ct]] = ct_numbering[[ct]] + 1
+}
+
+celltypes = as.list(celltypes)
 if (is.null(newcol)) {
-    sobj$seurat_clusters_id = sobj$seurat_clusters
-    celltypes$object = sobj
-    sobj = do_call(RenameIdents, celltypes)
-    sobj$seurat_clusters = Idents(sobj)
+    sobj <- rename_idents(sobj, ident, celltypes)
 } else {
-    celltypes$object = sobj
-    sobj = do_call(RenameIdents, celltypes)
-    sobj[[newcol]] = Idents(sobj)
-    Idents(sobj) = "seurat_clusters"
+    sobj@meta.data[[newcol]] = celltypes[as.character(Idents(sobj))]
 }
-
-print("- Saving Seurat object...")
-saveRDS(sobj, outfile)
-
-print("- Saving the mappings ...")
 celltypes$object = NULL
+gc()
+
+if (merge_same_labels) {
+    log$info("Merging clusters with the same labels...")
+    sobj <- merge_clusters_with_same_labels(sobj, newcol)
+    celltypes <- lapply(celltypes, function(ct) {
+        sub("\\.\\d+$", "", ct)
+    })
+}
+
+log$info("Saving the mappings ...")
 write.table(
     data.frame(
         Cluster = names(celltypes),
@@ -126,3 +146,6 @@ write.table(
     quote = FALSE,
     row.names = FALSE
 )
+
+log$info("Saving Seurat object...")
+biopipen.utils::save_obj(sobj, outfile)

biopipen/scripts/scrna/CellTypeAnnotation.R

@@ -1,13 +1,49 @@
+library(Seurat)
+library(biopipen.utils)
 set.seed(8525)
 
+backup_col <- {{envs.backup_col | r}}
+
+
+merge_clusters_with_same_labels <- function(sobj, newcol = NULL) {
+    if (is.null(newcol)) {
+        newcol <- biopipen.utils::GetIdentityColumn(sobj)
+        sobj@meta.data[[newcol]] <- sub("\\.\\d+$", "", sobj@meta.data[[newcol]])
+        Idents(sobj) <- newcol
+    } else {
+        sobj@meta.data[[newcol]] <- sub("\\.\\d+$", "", sobj@meta.data[[newcol]])
+    }
+
+    sobj
+}
+
+rename_idents <- function(sobj, ident_col, mapping) {
+    orig_ident_col <- biopipen.utils::GetIdentityColumn(sobj)
+    if (!identical(ident_col, orig_ident_col)) {
+        Idents(sobj) <- ident_col
+        mapping$object <- sobj
+        sobj <- do_call(RenameIdents, mapping)
+    } else {
+        if (!is.null(backup_col)) {
+            sobj@meta.data[[backup_col]] <- Idents(sobj)
+        }
+        mapping$object <- sobj
+        sobj <- do_call(RenameIdents, mapping)
+    }
+    sobj@meta.data[[ident_col]] <- Idents(sobj)
+    sobj
+}
+
 {% if envs.tool == "hitype" %}
 {% include biopipen_dir + "/scripts/scrna/CellTypeAnnotation-hitype.R" %}
 {% elif envs.tool == "sctype" %}
 {% include biopipen_dir + "/scripts/scrna/CellTypeAnnotation-sctype.R" %}
 {% elif envs.tool == "sccatch" %}
 {% include biopipen_dir + "/scripts/scrna/CellTypeAnnotation-sccatch.R" %}
+{% elif envs.tool == "celltypist" %}
+{% include biopipen_dir + "/scripts/scrna/CellTypeAnnotation-celltypist.R" %}
 {% elif envs.tool == "direct" %}
 {% include biopipen_dir + "/scripts/scrna/CellTypeAnnotation-direct.R" %}
 {% else %}
-stop(paste0("Unknown tool: ", {{envs.tool}}))
+stop("Unknown tool: {{envs.tool}}")
 {% endif %}