added a plotting function for reduced dim

.gitignore

@@ -30,4 +30,5 @@ renv/settings.json
 .Rprofile
 tidybulk.Rcheck/*
 ..Rcheck/*
-vignettes/cache/*
+vignettes/cache/*
+tidybulk_check.Rproj

NAMESPACE

@@ -22,6 +22,7 @@ export(log10_reverse_trans)
 export(logit_trans)
 export(pivot_sample)
 export(pivot_transcript)
+export(plot_reduced_dim)
 export(quantile_normalise_abundance)
 export(reduce_dimensions)
 export(remove_redundancy)

R/plotting_functions.R

@@ -0,0 +1,61 @@
+#' Plot reduced dimensions
+#'
+#' @param SE A 'SummarizedExperiment'
+#' @param VAR A Symbol. What variable in colData to be used for color coding
+#' @param method A character string. IndicatingMethod of dimensionality reduction to be ploted 
+#' @param dims Vector of dimensions to be plotted
+#'
+#' @returns A ggplot object
+#' @export
+#'
+#' @examples
+#' ## Load airway dataset for examples
+#' data('airway', package = 'airway')
+#'   # Ensure a 'condition' column exists for examples expecting it
+#'
+#'     SummarizedExperiment::colData(airway)$condition <- SummarizedExperiment::colData(airway)$dex
+#'
+#'
+#' counts.PCA =
+#'  airway |>
+#'  identify_abundant() |>
+#'  reduce_dimensions(assay = "counts", method="PCA", .dims = 3)
+#'  
+#'  plot_reduced_dim(counts.PCA, condition, "PCA", 1:2)
+plot_reduced_dim <- function(SE, VAR, method = "PCA", dims = 1:2){
+  varname <- rlang::ensym(VAR)
+
+  dim_type <- dplyr::case_when(
+    stringr::str_to_lower(method) == stringr::str_to_lower("PCA") ~ "pca",
+    stringr::str_to_lower(method) == stringr::str_to_lower("MDS") ~ "mds",
+    stringr::str_to_lower(method) == stringr::str_to_lower("tSNE") ~ "tsne",
+    stringr::str_to_lower(method) == stringr::str_to_lower("UMAP") ~ "umap")
+
+  coord_naming <- dplyr::case_when(
+    dim_type == "pca" ~ "PC",
+    dim_type == "mds" ~ "Dim",
+    dim_type == "tsne" ~ "tSNE",
+    dim_type == "umap" ~ "UMAP")
+
+  df <- SE %>% 
+    pivot_sample()
+
+  coord_names <- names(df)[stringr::str_detect(names(df), paste0("^", coord_naming, "\\d+"))]
+  coord_names <- coord_names[dims]
+
+
+  p <- SE |>
+    pivot_sample() |>
+    ggplot2::ggplot(ggplot2::aes(.data[[coord_names[1]]], .data[[coord_names[2]]], color = !!varname))+
+    ggplot2::geom_point()
+
+  if(dim_type == "pca"){
+    pca_res <- S4Vectors::metadata(SE)$tidybulk$PCA
+    vars <- pca_res[[1]]^2 / sum(pca_res[[1]]^2)
+
+    p <- p +
+      ggplot2::labs(x = paste0("PC1 (", round(vars[1] * 100), "% variance explained)"),
+                    y = paste0("PC2 (", round(vars[2] * 100), "% variance explained)"))
+  }
+  return(p)
+}