scripts: Add sliding GSEA for publication

scripts/sliding_gsea_package.R

@@ -8,7 +8,7 @@ bucket_size <- 500
 library(data.table)
 library(here)
 
-i_am("scripts/gsea.R")
+i_am("scripts/sliding_gsea_package.R")
 file_path <- here("scripts", "output", "ubigen_gsea.Rds")
 image_path <- here("scripts", "output", "ubigen_gsea.svg")
 

scripts/sliding_gsea_publication.R

@@ -0,0 +1,133 @@
+library(data.table)
+library(here)
+
+i_am("scripts/sliding_gsea_publication.R")
+
+gsea_sources <- c(
+  "GO:MF",
+  "GO:BP",
+  "GO:CC",
+  "KEGG",
+  "REAC",
+  "WP",
+  "TF",
+  "MIRNA",
+  "HPA",
+  "CORUM",
+  "HP"
+)
+
+gsea_source_colors <- data.table(
+  source = gsea_sources,
+  color = c(
+    "#dc3912",
+    "#ff9900",
+    "#109618",
+    "#dd4477",
+    "#3366cc",
+    "#0099c6",
+    "#5574a6",
+    "#22aa99",
+    "#6633cc",
+    "#66aa00",
+    "#990099"
+  )
+)
+
+sliding_gsea <- function(ranking, bucket_size_percent = 2.5) {
+  data <- copy(ranking)
+  data[, bucket := floor(percentile * 100 / bucket_size_percent)]
+  data[, bucket := (bucket * bucket_size_percent) / 100]
+
+  result <- data[, .(analysis = list(gprofiler2::gost(
+    gene,
+    custom_bg = ranking$gene,
+    domain_scope = "custom_annotated"
+  ))), by = bucket]
+
+  result[, result := lapply(analysis, function(a) a$result)]
+  result <- result[, rbindlist(result), by = bucket]
+
+  result[, .(count = .N), by = c("bucket", "source")]
+}
+
+sliding_gsea_plot <- function(
+    gsea_data,
+    sources = c("GO:MF", "GO:BP", "GO:CC", "KEGG", "REAC", "WP", "TF", "HP")) {
+  data <- merge(
+    gsea_data,
+    gsea_source_colors[source %chin% sources],
+    by = "source"
+  )
+
+  data[, source := factor(
+    source,
+    levels = gsea_source_colors[, source]
+  )]
+
+  setorder(data, source, bucket)
+
+  plotly::plot_ly() |>
+    plotly::add_bars(
+      data = data,
+      x = ~bucket,
+      y = ~count,
+      color = ~source,
+      marker = ~ list(color = color)
+    ) |>
+    plotly::layout(
+      xaxis = list(
+        title = "Bucket of genes (Percentile)",
+        tickformat = ".0%"
+      ),
+      yaxis = list(title = "Number of associated terms"),
+      barmode = "stack",
+      font = list(size = 8),
+      margin = list(
+        pad = 2,
+        l = 0,
+        r = 0,
+        t = 0,
+        b = 0
+      )
+    )
+}
+
+data_gtex <- sliding_gsea(ubigen::rank_genes(ubigen::gtex_all))
+data_cmap <- sliding_gsea(ubigen::rank_genes(ubigen::cmap))
+
+fig_gtex <- sliding_gsea_plot(data_gtex)
+fig_cmap <- sliding_gsea_plot(data_cmap)
+
+# Plotly specifies all sizes in pixels, including font size. Because of
+# that, we can actually think of these pixels as points. One point is defined as
+# 1/72 inch and SVG uses 96 DPI as the standard resolution.
+#
+# 1 plotly pixel = 1 point = 1/72 inch = 1 1/3 actual pixels
+#
+# So, we specify width and height in points (= plotly pixels) and scale up the
+# image by 96/72 to convert everything from points to pixels at 96 DPI.
+
+plotly::save_image(
+  fig_gtex |> plotly::hide_legend(),
+  file = here("scripts/output/sliding_gsea_gtex.svg"),
+  width = 6.27 * 72,
+  height = 3.135 * 72,
+  scale = 96 / 72
+)
+
+plotly::save_image(
+  fig_cmap |> plotly::hide_legend(),
+  file = here("scripts/output/sliding_gsea_cmap.svg"),
+  width = 6.27 * 72,
+  height = 3.135 * 72,
+  scale = 96 / 72
+)
+
+plotly::save_image(
+  fig_gtex,
+  file = here("scripts/output/sliding_gsea_legend.svg"),
+  width = 6.27 * 72,
+  height = 3.135 * 72,
+  scale = 96 / 72
+)