Merge pull request #397 from ncborcherding/dev

v2.0.5

DESCRIPTION

@@ -1,20 +1,21 @@
 Package: scRepertoire
 Title: A toolkit for single-cell immune receptor profiling
-Version: 2.0.4
+Version: 2.0.5
 Authors@R: c(
     person(given = "Nick", family = "Borcherding", role = c("aut", "cre"), email = "[email protected]"), 
-    person(given = "Qile", family = "Yang", role = c("aut"), email = "[email protected]"), 
+    person(given = "Qile", family = "Yang", role = c("aut"), email = "[email protected]"), 
     person(given = "Ksenia", family = "Safina", role = c("aut"), email = "[email protected]"))
 Description: scRepertoire is a toolkit for processing and analyzing single-cell T-cell receptor (TCR) and immunoglobulin (Ig). The scRepertoire framework supports use of 10x, AIRR, BD, MiXCR, Omniscope, TRUST4, and WAT3R single-cell formats. The functionality includes basic clonal analyses, repertoire summaries, distance-based clustering and interaction with the popular Seurat and SingleCellExperiment/Bioconductor R workflows.
 License: MIT + file LICENSE 
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 7.3.1
+RoxygenNote: 7.3.2
 biocViews: Software, ImmunoOncology, SingleCell, Classification, Annotation, Sequencing
 Depends: 
-	ggplot2, 
-	R (>= 4.0)
+     ggplot2, 
+     R (>= 4.0)
 Imports: 
+    assertthat,
     cubature,
     dplyr,
     evmix,
@@ -60,5 +61,5 @@ Config/testthat/edition: 3
 Language: en-US
 LinkingTo: 
     Rcpp
-URL: https://www.borch.dev/uploads/screpertoire/
+URL: https://www.borch.dev/uploads/scRepertoire/
 BugReports: https://github.com/ncborcherding/scRepertoire/issues

NAMESPACE

@@ -45,6 +45,9 @@ importFrom(SingleCellExperiment,reducedDim)
 importFrom(SummarizedExperiment,"colData<-")
 importFrom(SummarizedExperiment,colData)
 importFrom(VGAM,dpareto)
+importFrom(assertthat,assert_that)
+importFrom(assertthat,is.flag)
+importFrom(assertthat,is.string)
 importFrom(cubature,adaptIntegrate)
 importFrom(dplyr,"%>%")
 importFrom(dplyr,across)

NEWS.md

@@ -1,3 +1,10 @@
+# scRepertoire VERSION 2.0.5
+
+## UNDERLYING CHANGES
+* added type checks using assertthat
+* updated conditional statements in constructConDFAndparseTCR.cpp
+* Fixed issue in ```clonalQuant()``` and factor-based **group.by** variable
+
 # scRepertoire VERSION 2.0.4
 
 ## UNDERLYING CHANGES
@@ -8,7 +15,10 @@
 * ```clonalCompare()``` now retains the original clonal info if using **relabel.clones**
 * Add Dandelion support in to ```loadContigs()``` and testthat
 * Fixed issue with ```positionalProperty()``` assumption that the clones will all have 20 amino acids.
+* Fixed issue with ```positionalProperty()``` and removing non-amino acids.
 * Fixed IGH/K/L mistaking gene issue in ```vizGenes()```
+* Add error message for NULL results in ```clonalCluster()``` with **export.graph = TRUE**
+* Fixed issue with "full.clones" missing in ```combineExpression()``` when using 1 chain
 
 
 # scRepertoire VERSION 2.0.3

R/clonalCluster.R

@@ -130,23 +130,28 @@ clonalCluster <- function(input.data,
 
   #Returning the igraph object if exportGraph = TRUE
   if(exportGraph) {
-    cluster <- do.call(igraph::union, output.list)
-    vertex <- names(V(cluster))
-    data_df <- unique(data.frame(
-      id = vertex
-    ))
-    data_df <- merge(data_df, graph.variables, by = 1)
-    cluster <- set_vertex_attr(cluster, 
-                               name = "size", 
-                               index = data_df$id, 
-                               value = data_df[,2])
-    if(ncol(data_df) == 3) { #add grouping variable
+    if(length(is.null(output.list)) == length(output.list)) {
+      stop("No clusters detected with current parameters.")
+    } else {
+      output.list <- output.list[lapply(output.list,length)>0]
+      cluster <- do.call(igraph::union, output.list)
+      vertex <- names(V(cluster))
+      data_df <- unique(data.frame(
+        id = vertex
+      ))
+      data_df <- merge(data_df, graph.variables, by = 1)
       cluster <- set_vertex_attr(cluster, 
-                                 name = "group", 
+                                 name = "size", 
                                  index = data_df$id, 
-                                 value = data_df[,3])
+                                 value = data_df[,2])
+      if(ncol(data_df) == 3) { #add grouping variable
+        cluster <- set_vertex_attr(cluster, 
+                                   name = "group", 
+                                   index = data_df$id, 
+                                   value = data_df[,3])
+      }
+      return(cluster)
     }
-    return(cluster)
   }
 
   cluster.list <- lapply(seq_len(length(output.list)), function(x) {

R/clonalQuant.R

@@ -73,7 +73,7 @@ clonalQuant <- function(input.data,
       mat[i,3] <- length(na.omit(input.data[[i]][,cloneCall]))
       if (!is.null(group.by)) {
         location <- which(colnames(input.data[[i]]) == group.by)
-        mat[i,4] <- input.data[[i]][1,location]
+        mat[i,4] <- as.vector(input.data[[i]][1,location])
       }
   }
   if (scale) { 

R/combineContigs.R

@@ -47,6 +47,7 @@ utils::globalVariables(c(
 #' nonproductive chains if the variable exists in the contig data. Default
 #' is set to TRUE to remove nonproductive contigs.
 #' 
+#' @importFrom assertthat assert_that is.flag
 #' @export
 #' @concept Loading_and_Processing_Contigs
 #' @return List of clones for individual cell barcodes
@@ -58,6 +59,14 @@ combineTCR <- function(input.data,
                        removeMulti = FALSE, 
                        filterMulti = FALSE,
                        filterNonproductive = TRUE) {
+
+    # rudimentary input checking
+    assert_that(is.character(samples) || is.null(samples))
+    assert_that(is.character(ID) || is.null(ID))
+    assert_that(is.flag(removeNA))
+    assert_that(is.flag(removeMulti))
+    assert_that(is.flag(filterMulti))
+
     input.data <- .checkList(input.data)
     input.data <- .checkContigs(input.data)
     out <- NULL
@@ -185,6 +194,7 @@ combineTCR <- function(input.data,
 #' nonproductive chains if the variable exists in the contig data. Default
 #' is set to TRUE to remove nonproductive contigs.
 #' @importFrom dplyr %>% mutate
+#' @importFrom assertthat assert_that is.flag
 #' @export
 #' @concept Loading_and_Processing_Contigs
 #' @return List of clones for individual cell barcodes
@@ -200,6 +210,16 @@ combineBCR <- function(input.data,
     if(is.null(samples)) {
       stop("combineBCR() requires the samples paramter for the calculation of edit distance.")
     }
+
+    # rudimentary input checking
+    assert_that(is.character(samples) || is.null(samples))
+    assert_that(is.character(ID) || is.null(ID))
+    assert_that(is.flag(call.related.clones))
+    assert_that(is.numeric(threshold))
+    assert_that(is.flag(removeNA))
+    assert_that(is.flag(removeMulti))
+    assert_that(is.flag(filterMulti))
+
     input.data <- .checkList(input.data)
     input.data <- .checkContigs(input.data)
     out <- NULL

R/combineExpression.R

@@ -49,11 +49,12 @@
 #' @importFrom  rlang %||% sym :=
 #' @importFrom SummarizedExperiment colData<- colData
 #' @importFrom S4Vectors DataFrame
+#' @importFrom assertthat assert_that is.string is.flag
 #' @export
 #' @concept SC_Functions
 #' @return Single-cell object with clone information added to meta data
 #' information
-#' 
+#'
 combineExpression <- function(input.data, 
                               sc.data, 
                               cloneCall ="strict", 
@@ -64,6 +65,17 @@ combineExpression <- function(input.data,
                               cloneSize = c(Rare = 1e-4,Small = 0.001,Medium = 0.01,Large = 0.1,Hyperexpanded = 1),
                               addLabel = FALSE) {
     call_time <- Sys.time()
+
+    # rudimentary type checking
+    assert_that(isAnyValidProductOfCombineContigs(input.data))
+    assert_that(is_seurat_or_se_object(sc.data))
+    assert_that(is.string(cloneCall))
+    assert_that(is.string(chain))
+    assert_that(is.string(group.by) || is.null(group.by))
+    assert_that(is.flag(proportion))
+    assert_that(is.flag(filterNA))
+    assert_that(is_named_numeric(cloneSize))
+    assert_that(is.flag(addLabel))
 
     options( dplyr.summarise.inform = FALSE )
     if (!proportion && any(cloneSize < 1)) {
@@ -76,8 +88,9 @@ combineExpression <- function(input.data,
       #Retain the full clone information
       full.clone <- lapply(input.data, function(x) {
                         x[,c("barcode", cloneCall)]
-      full.clone <- bind_rows(full.clone)
+                 
       })
+      full.clone <- bind_rows(full.clone)
       for(i in seq_along(input.data)) {
         input.data[[i]] <- .off.the.chain(input.data[[i]], chain, cloneCall)
       }
@@ -179,7 +192,7 @@ combineExpression <- function(input.data,
       clone_sym <- sym(cloneCall)
       PreMeta <- PreMeta %>%
         left_join(full.clone, by = "barcode", suffix = c("", ".from_full_clones")) %>%
-        mutate(!!column_sym := coalesce(!!sym(paste0(cloneCall, ".from_full_clones")), !!column_sym)) %>%
+        mutate(!!clone_sym := coalesce(!!sym(paste0(cloneCall, ".from_full_clones")), !!clone_sym)) %>%
         select(-all_of(paste0(cloneCall, ".from_full_clones")))
     }
     barcodes <- PreMeta$barcode
@@ -226,14 +239,4 @@ combineExpression <- function(input.data,
     return(sc.data)
 } 
 
-
 .warn_str <- "< 1% of barcodes match: Ensure the barcodes in the single-cell object match the barcodes in the combined immune receptor output from scRepertoire. If getting this error, please check https://www.borch.dev/uploads/screpertoire/articles/faq."
-
-
-
-
-
-
-
-
-

R/positionalProperty.R

@@ -47,6 +47,7 @@
 #' @export
 #' @concept Summarize_Repertoire
 #' @return ggplot of line graph of diversity by position
+#' @author Florian Bach, Nick Borcherding
 
 positionalProperty <- function(input.data, 
                                chain = "TRB", 
@@ -83,6 +84,8 @@ positionalProperty <- function(input.data,
   #Getting AA Counts
   aa.count.list <- .aa.counter(input.data, cloneCall, aa.length)
 
+  aa.count.list <- lapply(aa.count.list, function(x)subset(x, x$AA %in% c("A", "R", "N", "D", "C", "Q", "E", "G", "H", "I", "L", "K", "M", "F", "P", "S", "T", "W", "Y", "V")))
+
   #Calculating properties and melting data
   lapply(seq_along(aa.count.list), function(x) {
       lapply(seq_len(nrow(aa.count.list[[x]]))[-1], function(y) {

R/typecheck.R

@@ -0,0 +1,56 @@
+# scRepertoire objects
+
+isCombineContigsOutput <- function(obj) {
+    is.list(obj) && all(sapply(obj, is.data.frame))
+}
+assertthat::on_failure(isCombineContigsOutput) <- function(call, env) {
+    paste0(deparse(call$obj), " is not an output of combineTCR or combineBCR")
+}
+
+isListOfTwoCombineContigsOutputs <- function(obj) {
+    is.list(obj) && length(obj) == 2 && all(sapply(obj, isCombineContigsOutput))
+}
+assertthat::on_failure(isListOfTwoCombineContigsOutputs) <- function(call, env) {
+    paste0(
+        deparse(call$obj),
+        " is not a list of two outputs of combineTCR and combineBCR"
+    )
+}
+
+isAnyValidProductOfCombineContigs <- function(obj) {
+    isCombineContigsOutput(obj) || isListOfTwoCombineContigsOutputs(obj)
+}
+assertthat::on_failure(isAnyValidProductOfCombineContigs) <- function(call, env) {
+    paste0(
+        deparse(call$obj),
+        " is not a valid output of combineTCR or combineBCR, nor a list of them"
+    )
+}
+
+# bio objects
+
+is_seurat_object <- function(obj) inherits(obj, "Seurat")
+assertthat::on_failure(is_seurat_object) <- function(call, env) {
+    paste0(deparse(call$obj), " is not a Seurat object")
+}
+
+is_se_object <- function(obj) inherits(obj, "SummarizedExperiment")
+assertthat::on_failure(is_se_object) <- function(call, env) {
+    paste0(deparse(call$obj), " is not a SummarizedExperiment object")
+}
+
+is_seurat_or_se_object <- function(obj) {
+    is_seurat_object(obj) || is_se_object(obj)
+}
+assertthat::on_failure(is_seurat_or_se_object) <- function(call, env) {
+    paste0(deparse(call$obj), " is not a Seurat or SummarizedExperiment object")
+}
+
+# general objects
+
+is_named_numeric <- function(obj) {
+    is.numeric(obj) && !is.null(names(obj))
+}
+assertthat::on_failure(is_named_numeric) <- function(call, env) {
+    paste0(deparse(call$obj), " is not a named numeric vector")
+}

R/utils.R

@@ -1,10 +1,5 @@
-# readability functions
+# readability functions with appropriate assertthat fail messages
 "%!in%" <- Negate("%in%")
-is_seurat_object <- function(obj) inherits(obj, "Seurat")
-is_se_object <- function(obj) inherits(obj, "SummarizedExperiment")
-is_seurat_or_se_object <- function(obj) {
-    is_seurat_object(obj) || is_se_object(obj)
-}
 
 #'@importFrom stringr str_sort
 .ordering.function <- function(vector, 
@@ -148,7 +143,7 @@ is_seurat_or_se_object <- function(obj) {
 .checkList <- function(df) {
   df <- tryCatch(
       {
-          if (!inherits(df, "list")) { 
+          if (!inherits(df, "list")) {
               df <- list(df)
           }
           df
@@ -629,7 +624,12 @@ is_df_or_list_of_df <- function(x) {
   edge.list <- do.call(rbind, edge.list)
 
   if(exportGraph) {
-    graph <- graph_from_edgelist(as.matrix(edge.list)[,c(1,2)])
+    if(!is.null(edge.list)) {
+      graph <- graph_from_edgelist(as.matrix(edge.list)[,c(1,2)])
+
+    } else {
+      graph <- NULL
+    }
     return(graph)
   }
 

_pkgdown.yml

@@ -47,6 +47,8 @@ navbar:
     - text: Quantifying Clonal Bias
       href: articles/Clonal_Bias.html
     - text: '-------'
+    - text: Making Deep Learning Models with immApex
+      href: articles/immApex.html
     - text: Combining Deep Learning and TCRs with Trex
       href: articles/Trex.html
     - text: Combining Deep Learning and BCRs with Ibex

inst/pkgdown.yml

@@ -1,5 +1,5 @@
 pandoc: 3.1.1
-pkgdown: 2.0.9
+pkgdown: 2.0.7
 pkgdown_sha: ~
 articles:
   Attaching_SC: Attaching_SC.html
@@ -18,8 +18,9 @@ articles:
   Running_Escape: Running_Escape.html
   SC_Visualizations: SC_Visualizations.html
   Trex: Trex.html
+  immApex: immApex.html
   vignette: vignette.html
-last_built: 2024-05-17T17:28Z
+last_built: 2024-07-15T18:40Z
 urls:
   reference: https://www.borch.dev/uploads/scRepertoire/reference
   article: https://www.borch.dev/uploads/scRepertoire/articles