Refactor prep (#81)

* some doc changes * let's start * Update plot_cpdb.R * start to simplify * Update test_cpdbplot1.R * Update plot_cpdb.R * Update utils.R * Update plot_cpdb.R tidy it up * Update plot_cpdb.R * Update plot_cpdb.R * Update misc.R * Update plot_cpdb.R * Update test_cpdbplot1.R * Update utils.R * change name * Update plot_cpdb.R * Update utils.R * Update plot_cpdb.R * ok let's mark it for now * hmm * Update plot_cpdb.R * Update plot_cpdb.R * Update utils.R * Update plot_cpdb.Rd * almost there. just plot_cpdb2 now * add vignette and fix functions * Update vignette.yml * fix bug * Update vignette.rmd * update readme * update options * add v5 compatibility * rename file * Update cpdb_output_v5.RData * hmm * Update plot_cpdb.R * Update cpdb_output_v5.RData * hmm * Update vignette_v5.rmd * Update cpdb_output_v5.RData * Update kidneyimmune.Rd * Update utils.R * Update plot_cpdb.R * Update plot_cpdb.R * make it look nicer * Update plot_cpdb.R * Update plot_cpdb.R * Update plot_cpdb.R * ooh the squish * Update vignette_v5.rmd * Update vignette_v5.rmd * update figure size and legend scale. * Update plot_cpdb.R * ok * Update codecov.yml * Update test_cpdbplot1.R * Update DESCRIPTION * ok make this work? * Update utils.R * Update plot_cpdb.R * Update plot_cpdb.R * Update utils.R * Update plot_cpdb.R * Update plot_cpdb.R * Update plot_cpdb.R * Update plot_cpdb.R * Update vignette_v5.rmd * Update vignette_v5.rmd * Update vignette_v5.rmd * update readme * Update vignette_v5.rmd * Update plot_cpdb.R
zktuong · Sep 12, 2023 · b70d98b · b70d98b
1 parent 1bd30b5
commit b70d98b
Show file tree

Hide file tree

Showing 32 changed files with 3,206 additions and 2,421 deletions.
diff --git a/.github/workflows/vignette.yml b/.github/workflows/vignette.yml
@@ -0,0 +1,68 @@
+on:
+  push:
+    branches:
+      - "master"
+  pull_request:
+    branches:
+      - "*"
+
+name: vignette
+
+jobs:
+  vignette:
+    defaults:
+      run:
+        shell: bash -l {0}
+    strategy:
+      matrix:
+        config:
+          - { os: macos-latest }
+    runs-on: ${{ matrix.config.os }}
+    env:
+      GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
+    steps:
+      - uses: actions/checkout@v3
+
+      - uses: r-lib/actions/setup-r@v2
+        id: install-r
+
+      - uses: r-lib/actions/setup-pandoc@v2
+
+      - id: bioc
+        name: Check bioc version
+        run: |
+          echo "##[set-output name=mainbiocversion;]$(Rscript -e 'cat(unlist(tools:::.BioC_version_associated_with_R_version()))' | awk '{print $1}')"
+          echo "##[set-output name=subbiocversion;]$(Rscript -e 'cat(unlist(tools:::.BioC_version_associated_with_R_version()))' | awk '{print $2}')"
+          echo "##[set-output name=biocversion;]$(Rscript -e 'cat(as.character(tools:::.BioC_version_associated_with_R_version()))' | awk '{print $1}')"
+        shell: bash -l {0}
+
+      - name: Install pak and query dependencies
+        run: |
+          install.packages("pak", repos = "https://r-lib.github.io/p/pak/dev/")
+          saveRDS(pak::pkg_deps("local::.", dependencies = TRUE), ".github/r-depends.rds")
+        shell: Rscript {0}
+
+      - name: Cache R packages
+        uses: actions/cache@v2
+        with:
+          path: |
+            ${{ env.R_LIBS_USER }}
+            !${{ env.R_LIBS_USER }}/pak
+          key: ${{ env.cache-version }}-${{ matrix.config.os }}-biocversion-RELEASE_${{ steps.bioc.outputs.mainbiocversion}}_${{ steps.bioc.outputs.subbiocversion}}-r-${{ matrix.config.r-version}}-${{ hashFiles('.github/depends.Rds') }}
+          restore-keys: ${{ env.cache-version }}-${{ matrix.config.os }}-biocversion-RELEASE_${{ steps.bioc.outputs.mainbiocversion}}_${{ steps.bioc.outputs.subbiocversion}}-r-${{ matrix.config.r-version}}-
+
+      - name: Install dependencies
+        run: |
+          pak::local_install_dev_deps(upgrade = TRUE, dependencies = c("all", "Config/Needs/website"))
+          pak::pkg_install("pkgdown")
+        shell: Rscript {0}
+
+      - name: Install package
+        run: R CMD INSTALL .
+
+      - name: Build and deploy pkgdown site
+        run: |
+          git config --local user.name "$GITHUB_ACTOR"
+          git config --local user.email "[email protected]"
+          Rscript -e 'pkgdown::deploy_to_branch(new_process = FALSE)'
+        shell: bash -l {0}
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,35 +1,33 @@
 Package: ktplots
 Title: Plot single-cell data dotplots
-Version: 1.2.5
+Version: 2.0.0
 Authors@R: person("Kelvin", "Tuong", email = c("[email protected]"), role = c("aut", "cre"))
 Description: Plotting tools for scData.
-License: MIT
+License: MIT + file LICENSE
 Encoding: UTF-8
 biocViews:
 Imports:
-	devtools,
-	Matrix,
-	dplyr,
-	ggplot2,
-	ggraph,
-	ggrepel,
-	reshape2,
-	viridis,
-	gtools,	
-	RColorBrewer,
-	purrr,
-	circlize,
-	ComplexHeatmap,
-	pheatmap
+    devtools,
+    Matrix,
+    dplyr,
+    ggplot2,
+    ggraph,
+    ggrepel,
+    reshape2,
+    viridis,
+    gtools,
+    RColorBrewer,
+    purrr,
+    circlize,
+    ComplexHeatmap,
+    pheatmap
 Suggests:
 	SummarizedExperiment,
 	SingleCellExperiment,
 	scater,
 	testthat,
 	covr,
 	plyr
-Depends:
-	ggplot2
 LazyData: true
 RoxygenNote: 7.2.3
 Collate: 

diff --git a/R/combine_cpdb.R b/R/combine_cpdb.R
@@ -13,19 +13,17 @@
 #' @export
 combine_cpdb <- function(...) {
     output <- list(...)
-    anames <- c(
-        "id_cp_interaction", "interacting_pair", "partner_a", "partner_b",
+    anames <- c("id_cp_interaction", "interacting_pair", "partner_a", "partner_b",
         "gene_a", "gene_b", "secreted", "receptor_a", "receptor_b", "annotation_strategy",
-        "is_integrin"
-    )
-    bnames <- c(
-        "gene_name", "uniprot", "is_complex", "protein_name", "complex_name",
-        "id_cp_interaction"
-    )
+        "is_integrin")
+    bnames <- c("gene_name", "uniprot", "is_complex", "protein_name", "complex_name",
+        "id_cp_interaction")
     if (all(colnames(output[[1]])[1:11] == anames)) {
-        out <- output %>% reduce(full_join, by = anames)
+        out <- output %>%
+            reduce(full_join, by = anames)
     } else if (all(colnames(output[[1]])[1:6] == bnames)) {
-        out <- output %>% reduce(full_join, by = bnames)
+        out <- output %>%
+            reduce(full_join, by = bnames)
     }
     return(out)
-}
+}
diff --git a/R/correlationSpot.R b/R/correlationSpot.R
@@ -20,22 +20,10 @@
 #' @return SpatialFeaturePlot
 #' @export
 
-correlationSpot <- function(
-    st,
-    genes = NULL,
-    celltypes = NULL,
-    geneset = NULL,
-    mode = c("high", "low", "both"),
-    cutoff = 0.5,
-    standardize = TRUE,
-    dims = 1:30,
-    k.params = 10,
-    resolution = 1,
-    rna_slot = "SCT",
-    label_slot = "predictions",
-    by = c("image", "expression"),
-    average_by_cluster = FALSE,
-    ...) {
+correlationSpot <- function(st, genes = NULL, celltypes = NULL, geneset = NULL, mode = c("high",
+    "low", "both"), cutoff = 0.5, standardize = TRUE, dims = 1:30, k.params = 10,
+    resolution = 1, rna_slot = "SCT", label_slot = "predictions", by = c("image",
+        "expression"), average_by_cluster = FALSE, ...) {
     requireNamespace("Seurat")
     requireNamespace("FNN")
 
@@ -44,7 +32,7 @@ correlationSpot <- function(
     st1 <- Seurat::FindNeighbors(st1, reduction = "pca", dims = dims, k.param = k.params)
     if (average_by_cluster) {
         st1 <- Seurat::FindClusters(st1, verbose = FALSE, resolution = resolution)
-        st1 <- Seurat::AddMetaData(st1, Idents(st1), "main_cluster")
+        st1 <- Seurat::AddMetaData(st1, Seurat::Idents(st1), "main_cluster")
     }
 
     # extract SNN graph
@@ -53,7 +41,8 @@ correlationSpot <- function(
         knn_idx <- FNN::get.knn(graph, k = k.params)$nn.index
     } else if (by == "image") {
         mat <- st1@images$slice1@coordinates[, c("row", "col")]
-        graph <- as.matrix(dist(mat))
+        requireNamespace("stats")
+        graph <- as.matrix(stats::dist(mat))
         knn_idx <- FNN::get.knn(graph, k = k.params)$nn.index
     }
 
@@ -109,7 +98,8 @@ correlationSpot <- function(
     }
 
     # get the nth percentile of each column
-    r1 <- apply(df, 2, quantile, cutoff)
+    requireNamespace("stats")
+    r1 <- apply(df, 2, stats::quantile, cutoff)
     if (mode != "both") {
         if (mode == "high") {
             for (i in seq_along(r1)) {
@@ -126,28 +116,29 @@ correlationSpot <- function(
         if (ncol(df) == 2) {
             df <- cbind(as.data.frame(df), main_cluster = st1@meta.data[, c("main_cluster")])
         } else {
+            requireNamespace("stats")
             if (length(genes) > 1 && length(celltypes) > 1) {
-                dx_g <- prcomp(df[, genes])
-                dx_c <- prcomp(df[, celltypes])
+                dx_g <- stats::prcomp(df[, genes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- cbind(dx_g, dx_c, main_cluster = st1@meta.data[, "main_cluster"])
             } else if (length(genes) > 1 && length(celltypes) < 1) {
-                dx_g <- prcomp(df[, genes])
+                dx_g <- stats::prcomp(df[, genes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 df <- cbind(dx_g, main_cluster = st1@meta.data[, "main_cluster"])
             } else if (length(celltypes) > 1 && length(genes) == 1) {
                 dx_g <- df[, genes]
-                dx_c <- prcomp(df[, celltypes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- cbind(dx_g, dx_c, main_cluster = st1@meta.data[, c("main_cluster")])
             } else if (length(celltypes) == 1 && length(genes) > 1) {
                 dx_c <- df[, celltypes]
-                dx_g <- prcomp(df[, genes])
+                dx_g <- stats::prcomp(df[, genes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 df <- cbind(dx_g, dx_c, main_cluster = st1@meta.data[, c("main_cluster")])
             } else if (length(celltypes) > 1 && length(genes) < 1) {
-                dx_c <- prcomp(df[, celltypes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- cbind(dx_c, main_cluster = st1@meta.data[, "main_cluster"])
             }
@@ -156,28 +147,29 @@ correlationSpot <- function(
         if (ncol(df) == 2) {
             df <- as.data.frame(df)
         } else {
+            requireNamespace("stats")
             if (length(genes) > 1 && length(celltypes) > 1) {
-                dx_g <- prcomp(df[, genes])
-                dx_c <- prcomp(df[, celltypes])
+                dx_g <- stats::prcomp(df[, genes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- cbind(dx_g, dx_c)
             } else if (length(celltypes) > 1 && length(genes) == 1) {
                 dx_g <- df[, genes]
-                dx_c <- prcomp(df[, celltypes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- cbind(dx_g, dx_c)
             } else if (length(celltypes) == 1 && length(genes) > 1) {
                 dx_c <- dx_g[, celltypes]
-                dx_g <- prcomp(df[, genes])
+                dx_g <- stats::prcomp(df[, genes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 df <- cbind(dx_g, dx_c)
             } else if (length(genes) > 1 && length(celltypes) < 1) {
-                dx_g <- prcomp(df[, genes])
+                dx_g <- stats::prcomp(df[, genes])
                 dx_g <- as.data.frame(dx_g$x)[, 1]
                 df <- dx_g
             } else if (length(celltypes) > 1 && length(genes) < 1) {
-                dx_c <- prcomp(df[, celltypes])
+                dx_c <- stats::prcomp(df[, celltypes])
                 dx_c <- as.data.frame(dx_c$x)[, 1]
                 df <- dx_c
             }
@@ -193,13 +185,15 @@ correlationSpot <- function(
 
     if (average_by_cluster) {
         nn_list <- lapply(nn_list, function(x) {
+            requireNamespace("stats")
             tmp <- x[, -which(colnames(x) %in% c("main_cluster"))]
-            res <- cor(tmp[, 1], tmp[, 2])
+            res <- stats::cor(tmp[, 1], tmp[, 2])
             return(res)
         })
     } else {
         nn_list <- lapply(nn_list, function(tmp) {
-            res <- cor(tmp[, 1], tmp[, 2])
+            requireNamespace("stats")
+            res <- stats::cor(tmp[, 1], tmp[, 2])
             return(res)
         })
     }
@@ -219,4 +213,4 @@ correlationSpot <- function(
 
     g <- Seurat::SpatialFeaturePlot(st, features = "correlation", ...)
     return(g)
-}
+}
diff --git a/R/data.R b/R/data.R
@@ -108,3 +108,53 @@
 #' @rdname kidneyimmune
 #' @docType data
 "covid_sample_metadata"
+
+#' cpdb_output_v5
+#'
+#' cpdb_output_v5 - Dataframe of CellPhoneDB output means.txt file
+#' @rdname kidneyimmune
+#' @docType data
+#' @usage data(cpdb_output_v5)
+#' @format data after CellPhoneDB v5 analysis
+#' @examples
+#' data(cpdb_output_v5)
+"means_v5"
+
+
+#' sce_v5
+#'
+#' sce_v5 - A small dummy singlecelldata for cellphonedb v5
+#' @rdname kidneyimmune
+#' @docType data
+"sce_v5"
+
+
+#' relevant_interactions_v5
+#'
+#' relevant_interactions_v5 - Dataframe of CellPhoneDB output relevant_interactions.txt file
+#' @rdname kidneyimmune
+#' @docType data
+"relevant_interactions_v5"
+
+
+#' decon_v5
+#'
+#' decon_v5 - Dataframe of CellPhoneDB output deconvoluted.txt file
+#' @rdname kidneyimmune
+#' @docType data
+"decon_v5"
+
+
+#' cellsign_v5
+#'
+#' cellsign_v5 - Dataframe of CellPhoneDB output CellSign.txt file
+#' @rdname kidneyimmune
+#' @docType data
+"cellsign_v5"
+
+#' interaction_scores_v5
+#'
+#' interaction_scores_v5 - Dataframe of CellPhoneDB output interaction_scores.txt file
+#' @rdname kidneyimmune
+#' @docType data
+"interaction_scores_v5"