Merge pull request #108 from jr-leary7/dev

Dev

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: scLANE
 Type: Package
 Title: Model gene expression dynamics with spline-based NB GLMs, GEEs, & GLMMs
-Version: 0.7.0
+Version: 0.7.1
 Authors@R: c(person(given = "Jack", family = "Leary", email = "j.leary@ufl.edu", role = c("aut", "cre")), 
              person(given = "Rhonda", family = "Bacher", email = "rbacher@ufl.edu", role = c("ctb", "fnd")))
 Description: This package uses truncated power basis spline models to build flexible, interpretable models of single cell gene expression over pseudotime or latent time. 

NEWS.md

@@ -1,3 +1,8 @@
+# scLANE 0.7.1
+
+* Changed input format of all functions to allow counts matrices formatted as `SingleCellExperiment` or `Seurat` objects, sparse matrices, or dense matrices.
+* Updated visualization functions to reflect changes made in `ggplot2` v3.4 (mostly changing the `size` parameter in line-based geoms to be `linewidth` instead). 
+
 # scLANE 0.6.3
 
 * Added a `NEWS.md` file to track changes to the package.

R/createCellOffset.R

@@ -3,7 +3,7 @@
 #' @name createCellOffset
 #' @author Jack Leary
 #' @description Creates a vector of per-cell size factors to be used as input to \code{\link{testDynamic}} as a model offset given a variety of inputs.
-#' @param expr.mat Either a gene-by-cell (cells as columns) matrix of raw integer counts prior to any cell filtering, a \code{Seurat} object, or a \code{SingleCellExperiment} object. Defaults to NULL.
+#' @param expr.mat Either a matrix of raw integer counts (cells as columns), a \code{Seurat} object, or a \code{SingleCellExperiment} object. Defaults to NULL.
 #' @param scale.factor The scaling factor use to multiply the sequencing depth factor for each cell. The default value is 1e4, which returns counts-per-10k.
 #' @return A named numeric vector containing the computed size factor for each cell.
 #' @seealso \code{\link{testDynamic}}
@@ -13,8 +13,8 @@
 #' @export
 #' @examples
 #' \dontrun{
-#' createCellOffset(expr.mat = raw_counts)
-#' createCellOffset(expr.mat = raw_counts, scale.factor = 1e5)
+#' createCellOffset(expr.mat = counts(sce_obj))
+#' createCellOffset(expr.mat = seu_obj, scale.factor = 1e5)
 #' }
 
 createCellOffset <- function(expr.mat = NULL, scale.factor = 1e4) {
@@ -26,7 +26,10 @@ createCellOffset <- function(expr.mat = NULL, scale.factor = 1e4) {
     expr.mat <- as.matrix(Seurat::GetAssayData(expr.mat,
                                                slot = "counts",
                                                assay = Seurat::DefaultAssay(expr.mat)))
+  } else if (inherits(expr.mat, "dgCMatrix")) {
+    expr.mat <- as.matrix(expr.mat)
   }
+  if (!(inherits(expr.mat, "matrix") || inherits(expr.mat, "array"))) { stop("Input expr.mat must be coerceable to a matrix of integer counts.") }
   # compute per-cell size factors
   cell_names <- colnames(expr.mat)
   seq_depths <- colSums(expr.mat)

R/getFittedValues.R

@@ -10,7 +10,7 @@
 #' @param test.dyn.res The output from \code{\link{testDynamic}}. Defaults to NULL.
 #' @param genes A character vector of genes to generate fitted values for. Defaults to NULL.
 #' @param pt A data.frame of pseudotime values for each cell. Defaults to NULL.
-#' @param expr.mat A matrix of integer expression values for each cell & gene. Must have genes as columns & cells as rows, with column names being gene names. Defaults to NULL.
+#' @param expr.mat Either a \code{SingleCellExperiment} or \code{Seurat} object from which counts can be extracted, or a matrix of integer-valued counts with genes as rows & cells as columns. Defaults to NULL.
 #' @param size.factor.offset (Optional) An offset to be used to rescale the fitted values. Can be generated easily with \code{\link{createCellOffset}}. No need to provide if the GEE backend was used. Defaults to NULL.
 #' @param cell.meta.data (Optional) A data.frame of metadata values for each cell (celltype label, subject characteristics, tissue type, etc.) that will be included in the result table. Defaults to NULL.
 #' @param id.vec (Optional) A vector of subject IDs used in fitting GEE or GLMM models. Defaults to NULL.
@@ -23,8 +23,8 @@
 #' getFittedValues(gene_stats,
 #'                 genes = c("Neurog3", "Epcam", "Krt19"),
 #'                 pt = pt_df,
-#'                 expr.mat = gene_counts,
-#'                 cell.meta.data = seurat_object@meta.data,
+#'                 expr.mat = seu_obj,
+#'                 cell.meta.data = seu_obj@meta.data,
 #'                 ci.alpha = 0.05)
 #' }
 
@@ -39,6 +39,25 @@ getFittedValues <- function(test.dyn.res = NULL,
                             filter.lineage = NULL) {
   # check inputs
   if (is.null(expr.mat) || is.null(pt) || is.null(genes) || is.null(test.dyn.res)) { stop("You forgot one or more of the arguments to getFittedValues().") }
+  # get raw counts from SingleCellExperiment or Seurat object & transpose to cell x gene dense matrix
+  if (inherits(expr.mat, "SingleCellExperiment")) {
+    expr.mat <- BiocGenerics::counts(expr.mat)[genes, ]
+    expr.mat <- as.matrix(expr.mat)
+  } else if (inherits(expr.mat, "Seurat")) {
+    expr.mat <- Seurat::GetAssayData(expr.mat,
+                                     slot = "counts",
+                                     assay = Seurat::DefaultAssay(expr.mat))
+    expr.mat <- as.matrix(expr.mat[genes, ])
+  } else if (inherits(expr.mat, "dgCMatrix")) {
+    expr.mat <- as.matrix(expr.mat)
+  }
+  if (!(inherits(expr.mat, "matrix") || inherits(expr.mat, "array"))) { stop("Input expr.mat must be coerceable to a matrix of integer counts.") }
+  expr.mat <- t(expr.mat)  # transpose to cell x gene matrix
+  if (is.null(genes)) {
+    genes <- colnames(expr.mat)
+  } else {
+    expr.mat <- expr.mat[, genes]
+  }
   # generate parameters for CIs
   Z <- stats::qnorm(ci.alpha / 2, lower.tail = FALSE)
   # select sublist for gene of interest

R/plotModels.R

@@ -70,18 +70,23 @@ plotModels <- function(test.dyn.res = NULL,
                        plot.gam = TRUE,
                        plot.scLANE = TRUE,
                        filter.lineage = NULL,
-                       gg.theme = ggplot2::theme_classic(base_size = 14)) {
+                       gg.theme = ggplot2::theme_classic(base_size = 14,
+                                                         base_line_size = 0.75,
+                                                         base_rect_size = 0.75)) {
   # check inputs
   if (is.null(expr.mat) || is.null(pt) || is.null(gene) || is.null(test.dyn.res)) { stop("You forgot one or more of the arguments to plotModels().") }
   # get raw counts from SingleCellExperiment or Seurat object & transpose to cell x gene dense matrix
   if (inherits(expr.mat, "SingleCellExperiment")) {
-    expr.mat <- as.matrix(t(BiocGenerics::counts(expr.mat)))
+    expr.mat <- as.matrix(BiocGenerics::counts(expr.mat))
   } else if (inherits(expr.mat, "Seurat")) {
-    expr.mat <- as.matrix(t(Seurat::GetAssayData(expr.mat,
-                                                 slot = "counts",
-                                                 assay = Seurat::DefaultAssay(expr.mat))))
+    expr.mat <- as.matrix(Seurat::GetAssayData(expr.mat,
+                                               slot = "counts",
+                                               assay = Seurat::DefaultAssay(expr.mat)))
+  } else if (inherits(expr.mat, "dgCMatrix")) {
+    expr.mat <- as.matrix(expr.mat)
   }
   if (!(inherits(expr.mat, "matrix") || inherits(expr.mat, "array"))) { stop("Input expr.mat must be coerceable to a matrix of integer counts.") }
+  expr.mat <- t(expr.mat)
   # generate parameters for CIs
   Z <- stats::qnorm(ci.alpha / 2, lower.tail = FALSE)
   # select sublist for gene of interest
@@ -277,13 +282,13 @@ plotModels <- function(test.dyn.res = NULL,
     }
     p <- p +
          ggplot2::geom_line(mapping = ggplot2::aes(y = PRED),
-                            size = 0.75,
+                            linewidth = 0.75,
                             color = "black",
                             show.legend = FALSE) +
          ggplot2::geom_ribbon(mapping = ggplot2::aes(ymin = CI_LL, ymax = CI_UL),
                               fill = "grey50",
                               alpha = 0.4,
-                              size = 0,
+                              linewidth = 0,
                               show.legend = FALSE) +
          ggplot2::scale_y_continuous(labels = scales::label_comma()) +
          ggplot2::scale_x_continuous(labels = scales::label_number(accuracy = 0.1)) +
@@ -301,17 +306,17 @@ plotModels <- function(test.dyn.res = NULL,
                              size = 0.5,
                              show.legend = ifelse(ncol(pt) > 1, TRUE, FALSE))
     if (requireNamespace("ggh4x", quietly = TRUE)) {
-      p <- p + ggh4x::facet_nested_wrap(~paste0("Lineage ", LINEAGE) + MODEL)
+      p <- p + ggh4x::facet_nested_wrap(~paste0("Lineage ", LINEAGE) + MODEL, )
     } else {
       p <- p + ggplot2::facet_wrap(~paste0("Lineage ", LINEAGE) + MODEL)
     }
     p <- p +
          ggplot2::geom_line(mapping = ggplot2::aes(y = PRED),
-                            size = 0.75,
+                            linewidth = 0.75,
                             color = "black") +
          ggplot2::geom_ribbon(mapping = ggplot2::aes(ymin = CI_LL, ymax = CI_UL),
                               alpha = 0.4,
-                              size = 0,
+                              linewidth = 0,
                               color = "grey") +
          ggplot2::scale_y_continuous(labels = scales::label_comma()) +
          ggplot2::scale_x_continuous(labels = scales::label_number(accuracy = 0.1)) +

R/testDynamic.R

@@ -16,7 +16,7 @@
 #' @importFrom stats predict logLik deviance offset
 #' @importFrom geeM geem
 #' @importFrom glmmTMB glmmTMB nbinom2
-#' @param expr.mat Either a \code{SingleCellExperiment} or \code{Seurat} object from which counts can be extracted, or a dense matrix of integer-valued counts. Defaults to NULL.
+#' @param expr.mat Either a \code{SingleCellExperiment} or \code{Seurat} object from which counts can be extracted, or a matrix of integer-valued counts with genes as rows & cells as columns. Defaults to NULL.
 #' @param pt Either the output from \code{\link[slingshot]{SlingshotDataSet}} object from which pseudotime can be generated, or a data.frame containing the pseudotime or latent time estimates for each cell (can be multiple columns / lineages). Defaults to NULL.
 #' @param genes A character vector of genes to model. If not provided, defaults to all genes in \code{expr.mat}. Defaults to NULL.
 #' @param n.potential.basis.fns (Optional) The maximum number of possible basis functions. See the parameter \code{M} in \code{\link{marge2}}. Defaults to 5.
@@ -59,20 +59,20 @@
 #'             parallel.exec = TRUE,
 #'             n.cores = 8,
 #'             n.potential.basis.fns = 7)
-#' testDynamic(expr.mat = raw_counts,
+#' testDynamic(expr.mat = counts(sce_obj),
 #'             pt = pseudotime_df,
 #'             is.gee = TRUE,
-#'             id.vec = my_subject_ids,
+#'             id.vec = colData(sce_obj)$subject_id,
 #'             cor.structure = "ar1",
 #'             parallel.exec = TRUE,
 #'             n.cores = 8,
 #'             n.potential.basis.fns = 7)
-#' testDynamic(expr.mat = raw_counts,
+#' testDynamic(expr.mat = seu_obj,
 #'             pt = pseudotime_df,
 #'             parallel.exec = TRUE,
 #'             n.cores = 8,
 #'             is.glmm = TRUE,
-#'             id.vec = my_subject_ids,
+#'             id.vec = seu_obj$subject_id,
 #'             log.file = "scLANE_log.txt",
 #'             log.iter = 10)
 #' }
@@ -98,14 +98,17 @@ testDynamic <- function(expr.mat = NULL,
   # get raw counts from SingleCellExperiment or Seurat object & transpose to cell x gene dense matrix
   if (inherits(expr.mat, "SingleCellExperiment")) {
     expr.mat <- BiocGenerics::counts(expr.mat)[genes, ]
-    expr.mat <- t(as.matrix(expr.mat))
+    expr.mat <- as.matrix(expr.mat)
   } else if (inherits(expr.mat, "Seurat")) {
     expr.mat <- Seurat::GetAssayData(expr.mat,
                                      slot = "counts",
                                      assay = Seurat::DefaultAssay(expr.mat))
-    expr.mat <- t(as.matrix(expr.mat[genes, ]))
+    expr.mat <- as.matrix(expr.mat[genes, ])
+  } else if (inherits(expr.mat, "dgCMatrix")) {
+    expr.mat <- as.matrix(expr.mat)
   }
   if (!(inherits(expr.mat, "matrix") || inherits(expr.mat, "array"))) { stop("Input expr.mat must be coerceable to a matrix of integer counts.") }
+  expr.mat <- t(expr.mat)  # transpose to cell x gene matrix
   if (is.null(genes)) {
     genes <- colnames(expr.mat)
   } else {