Merge branch 'main-dev' into fix_factor_plot_colorscale

Author: jan-spr

DESCRIPTION

@@ -1,9 +1,9 @@
 Package: MOFA2
 Type: Package
 Title: Multi-Omics Factor Analysis v2
-Version: 1.7.2
-Maintainer: Britta Velten <britta.velten@gmail.com>
-Authors@R: c(person("Ricard", "Argelaguet", role = "aut",
+Version: 1.21.1
+Maintainer: Ricard Argelaguet <ricard.argelaguet@gmail.com>
+Authors@R: c(person("Ricard", "Argelaguet", role = c("aut", "cre"),
                      email = "[email protected]",
                      comment = c(ORCID = "http://orcid.org/0000-0003-3199-3722")),
               person("Damien", "Arnol", role = "aut",
@@ -12,23 +12,23 @@ Authors@R: c(person("Ricard", "Argelaguet", role = "aut",
               person("Danila", "Bredikhin", role = "aut",
                      email = "[email protected]",
                      comment = c(ORCID = "https://orcid.org/0000-0001-8089-6983")),                     
-              person("Britta", "Velten", role = c("aut", "cre"),
+              person("Britta", "Velten", role = "aut",
               		email = "[email protected]",
               		comment = c(ORCID = "http://orcid.org/0000-0002-8397-3515"))
               )
-Date: 2018-09-03
-License: LGPL-3.0
+Date: 2023-01-12
+License: file LICENSE
 Description: The MOFA2 package contains a collection of tools for training and analysing multi-omic factor analysis (MOFA). MOFA is a probabilistic factor model that aims to identify principal axes of variation from data sets that can comprise multiple omic layers and/or groups of samples. Additional time or space information on the samples can be incorporated using the MEFISTO framework, which is part of MOFA2. Downstream analysis functions to inspect molecular features underlying each factor, vizualisation, imputation etc are available.
 Encoding: UTF-8
 Depends: R (>= 4.0)
 Imports: rhdf5, dplyr, tidyr, reshape2, pheatmap, ggplot2, methods, RColorBrewer, cowplot, ggrepel, reticulate, HDF5Array, grDevices, stats, magrittr, forcats, utils, corrplot, DelayedArray, Rtsne, uwot, basilisk, stringi
-Suggests: knitr, testthat, Seurat, ggpubr, foreach, psych, MultiAssayExperiment, SummarizedExperiment, SingleCellExperiment, ggrastr, mvtnorm, GGally, rmarkdown, data.table, tidyverse, BiocStyle, Matrix, markdown
+Suggests: knitr, testthat, Seurat, SeuratObject, ggpubr, foreach, psych, MultiAssayExperiment, SummarizedExperiment, SingleCellExperiment, ggrastr, mvtnorm, GGally, rmarkdown, data.table, tidyverse, BiocStyle, Matrix, markdown
 biocViews: DimensionReduction, Bayesian, Visualization
 URL: https://biofam.github.io/MOFA2/index.html
 BugReports: https://github.com/bioFAM/MOFA2
 VignetteBuilder: knitr
 LazyData: false
 StagedInstall: no
 NeedsCompilation: yes
-RoxygenNote: 7.1.1
+RoxygenNote: 7.3.3
 SystemRequirements: Python (>=3), numpy, pandas, h5py, scipy, argparse, sklearn, mofapy2

NAMESPACE

@@ -134,7 +134,7 @@ importFrom(dplyr,left_join)
 importFrom(dplyr,mutate)
 importFrom(dplyr,summarise)
 importFrom(dplyr,top_n)
-importFrom(forcats,fct_explicit_na)
+importFrom(forcats,fct_na_value_to_level)
 importFrom(ggrepel,geom_text_repel)
 importFrom(grDevices,colorRampPalette)
 importFrom(magrittr,"%>%")

R/basilisk.R

@@ -1,21 +1,21 @@
 # .mofapy2_dependencies <- c(
 #     "h5py==3.1.0",
 #     "pandas==1.2.1",
-#   "scikit-learn==0.24.1",
-#   "dtw-python==1.1.10"
+#     "scikit-learn==0.24.1",
+#     "dtw-python==1.1.10"
 # )
 
 .mofapy2_dependencies <- c(
-    "python=3.10.5",
-    "numpy=1.23.1",
-    "scipy=1.8.1",
-    "pandas=1.4.3",
-    "h5py=3.6.0",
-    "scikit-learn=1.1.1",
-    "dtw-python=1.2.2"
+    "python=3.12.12",
+    "numpy=1.26.4",
+    "scipy=1.12.0",
+    "pandas=2.2.1",
+    "h5py=3.10.0",
+    "scikit-learn=1.4.0",
+    "dtw-python=1.3.1"
 )
 
-.mofapy2_version <- "0.6.6"
+.mofapy2_version <- "0.7.2"
 
 #' @importFrom basilisk BasiliskEnvironment
 mofa_env <- BasiliskEnvironment("mofa_env", pkgname="MOFA2", packages=.mofapy2_dependencies, pip = paste0("mofapy2==",.mofapy2_version))

R/calculate_variance_explained.R

@@ -277,8 +277,8 @@ plot_variance_explained <- function(object, x = "view", y = "factor", split_by =
 
 
   # Grid plot with the variance explained per factor and view/group
-  p1 <- ggplot(r2_mk_df, aes_string(x=x, y=y)) + 
-    geom_tile(aes_string(fill="value"), color="black") +
+  p1 <- ggplot(r2_mk_df, aes(x=.data[[x]], y=.data[[y]])) + 
+    geom_tile(aes(fill=.data$value), color="black") +
     facet_wrap(as.formula(sprintf('~%s',split_by)), nrow=1) +
     labs(x="", y="", title="") +
     scale_fill_gradientn(colors=c("gray97","darkblue"), guide="colorbar", limits=c(min_r2,max_r2)) +
@@ -313,21 +313,21 @@ plot_variance_explained <- function(object, x = "view", y = "factor", split_by =
     max_lim_bplt <- max(r2_m_df$R2)
 
     # Barplot with variance explained per view/group (across all factors)
-    p2 <- ggplot(r2_m_df, aes_string(x=x, y="R2")) + 
+    p2 <- ggplot(r2_m_df, aes(x=.data[[x]], y=.data$R2)) + 
       # ggtitle(sprintf("%s\nTotal variance explained per %s", i, x)) +
       geom_bar(stat="identity", fill="deepskyblue4", color="black", width=0.9) +
       facet_wrap(as.formula(sprintf('~%s',split_by)), nrow=1) +
       xlab("") + ylab("Variance explained (%)") +
       scale_y_continuous(limits=c(min_lim_bplt, max_lim_bplt), expand=c(0.005, 0.005)) +
       theme(
         axis.ticks.x = element_blank(),
-        axis.text.x = element_text(size=rel(1.1), color="black"),
-        axis.text.y = element_text(size=rel(1.0), color="black"),
-        axis.title.y = element_text(size=rel(1.0), color="black"),
-        axis.line = element_line(size=rel(1.0), color="black"),
+        axis.text.x = element_text(color="black"),
+        axis.text.y = element_text(color="black"),
+        axis.title.y = element_text(color="black"),
+        axis.line = element_line(color="black"),
         panel.background = element_blank(),
         strip.background = element_blank(),
-        strip.text = element_text(size=rel(1.0))
+        strip.text = element_text()
       )
 
     # remove facet title
@@ -515,8 +515,8 @@ plot_variance_explained_per_feature <- function(object, view, features = 10,
   }
 
   # Grid plot with the variance explained per feature in every group
-  p <- ggplot(r2_df, aes_string(x = "group", y = "feature")) + 
-    geom_tile(aes_string(fill = "value"), color = "black") +
+  p <- ggplot(r2_df, aes(x = .data$group, y = .data$feature)) + 
+    geom_tile(aes(fill = .data$value), color = "black") +
     guides(fill = guide_colorbar("R2 (%)")) +
     labs(x = "", y = "", title = "") +
     scale_fill_gradientn(colors=c("gray97","darkblue"), guide="colorbar", limits=c(min_r2, max_r2)) +

R/compare_models.R

@@ -121,7 +121,7 @@ compare_elbo <- function(models, log = FALSE, return_data = FALSE) {
   # return data
   if (return_data) return(df)
 
-  gg <- ggplot(df, aes_string(x="model", y="ELBO")) + 
+  gg <- ggplot(df, aes(x=.data$model, y=.data$ELBO)) + 
     geom_bar(stat="identity", color="black", fill="grey70") +
     labs(x="", y="Evidence Lower Bound (ELBO)") +
     theme_classic()

R/contribution_scores.R

@@ -111,7 +111,7 @@ plot_contribution_scores <- function(object, samples = "all", group_by = NULL, r
 
     to.plot <- scores
     if (return_data) return(to.plot)
-    p <- ggplot(to.plot, aes_string(x="view", y="value")) +
+    p <- ggplot(to.plot, aes(x=.data$view, y=.data$value)) +
       geom_bar(aes(fill=view), stat="identity", color="black") +
       facet_wrap(~sample) +
       labs(x="", y="Contribution score") +
@@ -129,7 +129,7 @@ plot_contribution_scores <- function(object, samples = "all", group_by = NULL, r
 
     to.plot <- merge(scores, object@samples_metadata[,c("sample",group_by)], by="sample")
     if (return_data) return(to.plot)
-    p <- ggplot(to.plot, aes_string(x="view", y="value")) +
+    p <- ggplot(to.plot, aes(x=.data$view, y=.data$value)) +
       geom_boxplot(aes(fill=view)) +
       facet_wrap(as.formula(paste("~", group_by))) +
       labs(x="", y="Contribution score") +

R/correlate_covariates.R

@@ -140,10 +140,10 @@ summarise_factors <- function(object, df, factors = "all", groups = "all", abs =
   # Plot
   if (length(unique(factors_df$group))>1) {
     to.plot$group <- factor(to.plot$group, levels=groups)
-    p <- ggplot(to.plot, aes_string(x="group", y="level", fill="value")) +
+    p <- ggplot(to.plot, aes(x=.data$group, y=.data$level, fill=.data$value)) +
       facet_wrap(~factor)
   } else {
-    p <- ggplot(to.plot, aes_string(x="factor", y="level", fill="value"))
+    p <- ggplot(to.plot, aes(x=.data$factor, y=.data$level, fill=.data$value))
   }
 
   p <- p +

R/create_mofa.R

@@ -370,21 +370,24 @@ create_mofa_from_SingleCellExperiment <- function(sce, groups = NULL, assay = "l
 #' Alternatively, a character vector with group assignment for every sample.
 #' Default is \code{NULL} (no group structure).
 #' @param assays assays to use, default is \code{NULL}, it fetched all assays available
-#' @param slot assay slot to be used (default is scale.data).
+#' @param layer layer to be used (default is data).
 #' @param features a list with vectors, which are used to subset features, with names corresponding to assays; a vector can be provided when only one assay is used
 #' @param extract_metadata logical indicating whether to incorporate the metadata from the Seurat object into the MOFA object
 #' @return Returns an untrained \code{\link{MOFA}} object
 #' @export
-create_mofa_from_Seurat <- function(seurat, groups = NULL, assays = NULL, slot = "scale.data", features = NULL, extract_metadata = FALSE) {
+create_mofa_from_Seurat <- function(seurat, groups = NULL, assays = NULL, layer = "data", features = NULL, extract_metadata = FALSE) {
 
   # Check is Seurat is installed
   if (!requireNamespace("Seurat", quietly = TRUE)) {
     stop("Package \"Seurat\" is required but is not installed.", call. = FALSE)
   } else {
 
+    # Check Seurat version
+    if (SeuratObject::Version(seurat)$major != 5) stop("Please install Seurat v5")
+    
     # Define assays
     if (is.null(assays)) {
-      assays <- Seurat::Assays(seurat)
+      assays <- SeuratObject::Assays(seurat)
       message(paste0("No assays specified, using all assays by default: ", paste(assays,collapse=" ")))
     } else {
       stopifnot(assays%in%Seurat::Assays(seurat))
@@ -429,7 +432,7 @@ create_mofa_from_Seurat <- function(seurat, groups = NULL, assays = NULL, slot =
 
     # Extract data matrices
     data_matrices <- lapply(assays, function(i) 
-      .split_seurat_into_groups(seurat, groups = groups, assay = i, slot = slot, features = features[[i]]))
+      .split_seurat_into_groups(seurat, groups = groups, assay = i, layer = layer, features = features[[i]]))
     names(data_matrices) <- assays
 
     # Create MOFA object
@@ -583,10 +586,10 @@ create_mofa_from_matrix <- function(data, groups = NULL) {
 }
 
 # (Hidden) function to split data in Seurat object into a list of matrices
-.split_seurat_into_groups <- function(seurat, groups, assay = "RNA", slot = "scale.data", features = NULL) {
-  data <- Seurat::GetAssayData(object = seurat, assay = assay, slot = slot)
+.split_seurat_into_groups <- function(seurat, groups, assay = "RNA", layer = "data", features = NULL) {
+  data <- SeuratObject::GetAssayData(object = seurat, assay = assay, layer = layer)
   if(is.null(data) | any(dim(data) == 0)){
-    stop(paste("No data present in the slot",slot, "of the assay",assay ,"in the Seurat object."))
+    stop(paste("No data present in the layer",layer, "of the assay",assay ,"in the Seurat object."))
   }
   if (!is.null(features)) data <- data[features, , drop=FALSE]
   .split_data_into_groups(list(data), groups)[[1]]

R/dimensionality_reduction.R

@@ -234,22 +234,22 @@ plot_dimred <- function(object, method = c("UMAP", "TSNE"), groups = "all", show
   if (is.null(stroke)) if (length(unique(df$sample))<1000) { stroke <- 0.5 } else { stroke <- 0 }
 
   # Generate plot
-  p <- ggplot(df, aes_string(x = "x", y = "y")) + 
+  p <- ggplot(df, aes(x = .data$x, y = .data$y)) + 
     labs(x = latent_dimensions_names[1], y = latent_dimensions_names[2]) +
     theme_classic() +
     theme(
       axis.text = element_blank(), 
       axis.title = element_blank(), 
-      axis.line = element_line(color = "black", size = 0.5), 
+      axis.line = element_line(color = "black", linewidth = 0.5), 
       axis.ticks = element_blank()
     )
 
   # Add dots  
   if (rasterize) {
     message("for rasterizing the plot we use ggrastr::geom_point_rast()")
-    p <- p + ggrastr::geom_point_rast(aes_string(fill = "color_by", shape = "shape_by", alpha = "observed"), size = dot_size, stroke = stroke)
+    p <- p + ggrastr::geom_point_rast(aes(fill = .data$color_by, shape = .data$shape_by, alpha = .data$observed), size = dot_size, stroke = stroke)
   } else {
-    p <- p + geom_point(aes_string(fill = "color_by", shape = "shape_by", alpha = "observed"), size = dot_size, stroke = stroke)
+    p <- p + geom_point(aes(fill = .data$color_by, shape = .data$shape_by, alpha = .data$observed), size = dot_size, stroke = stroke)
 
   }      
 
@@ -273,7 +273,7 @@ plot_dimred <- function(object, method = c("UMAP", "TSNE"), groups = "all", show
         return(data.medians)
       }
     ) %>% do.call("rbind",.)
-    p <- p + geom_text_repel(aes_string(label="color_by"), data=labels.loc)
+    p <- p + geom_text_repel(aes(label=.data$color_by), data=labels.loc)
   }
 
 

R/enrichment.R

@@ -267,11 +267,11 @@ plot_enrichment <- function(enrichment.results, factor, alpha = 0.1, max.pathway
   tmp$pathway <- factor(tmp$pathway <- rownames(tmp), levels = tmp$pathway[order(tmp$pvalue, decreasing = TRUE)])
   tmp$start <- 0
 
-  p <- ggplot(tmp, aes_string(x="pathway", y="logp")) +
+  p <- ggplot(tmp, aes(x=.data$pathway, y=.data$logp)) +
     geom_point(size=dot_size) +
     geom_hline(yintercept=-log10(alpha), linetype="longdash") +
     scale_color_manual(values=c("black","red")) +
-    geom_segment(aes_string(xend="pathway", yend="start")) +
+    geom_segment(aes(xend=.data$pathway, yend=.data$start)) +
     ylab("-log pvalue") +
     coord_flip() +
     theme(
@@ -406,10 +406,10 @@ plot_enrichment_detailed <- function(enrichment.results, factor,
   tmp$pathway_long_name <- factor(tmp$pathway_long_name, levels=order_pathways)
   tmp_filt$pathway_long_name <- factor(tmp_filt$pathway_long_name, levels=order_pathways)
 
-  p <- ggplot(tmp, aes_string(x="pathway_long_name", y="feature.statistic")) +
-    geom_text_repel(aes_string(x="pathway_long_name", y="feature.statistic", label="feature"), size=text_size, color="black", force=1, data=tmp_filt) +
+  p <- ggplot(tmp, aes(x=.data[["pathway_long_name"]], y=.data[["feature.statistic"]])) +
+    geom_text_repel(aes(x=.data[["pathway_long_name"]], y=.data[["feature.statistic"]], label=.data$feature), size=text_size, color="black", force=1, data=tmp_filt) +
     geom_point(size=0.5, color="lightgrey") +
-    geom_point(aes_string(x="pathway_long_name", y="feature.statistic"), size=1, color="black", data=tmp_filt) +
+    geom_point(aes(x=.data[["pathway_long_name"]], y=.data[["feature.statistic"]]), size=1, color="black", data=tmp_filt) +
     labs(x="", y="Weight (scaled)", title="") +
     coord_flip() +
     theme(