Merge pull request #3 from GaitiLab/dev

Dev

Author: joan-yanqiong

.gitignore

@@ -46,7 +46,6 @@ vignettes/*.pdf
 ._.*
 ._*
 .DS_Store
-jobs/pipeline_runs/*/.nextflow/*
 
 # Python
 Python/__pycache__
@@ -60,11 +59,19 @@ slurm_out
 
 output/
 data/
-misc/
+
 
 analysis/run_notebooks.R
 analysis/run_notebooks.sh
 TEST-*
 tmp.txt
 
-000_ARCHIVED
+**/000_ARCHIVED
+.nextflow
+preprocessing/scRNAseq/CCI_CellClass_L2_2/gaitilab.config
+
+
+misc/internal
+# TODO remove once supptables finalized
+misc/SuppTables
+internal_scripts

.lintr

@@ -6,5 +6,7 @@ linters: linters_with_defaults(
     indentation_linter(
         indent = 4L
     ),
-    commented_code_linter=NULL
+    commented_code_linter=NULL, 
+    line_length_linter(80L),
+    pipe_continuation_linter()
     )

README.md

@@ -1,11 +1,14 @@
 # Overview
 
-Repository with code to reproduce analyses and figures in "Glioblastoma cells co-opt epigenetic regulators of normal oligodendrocyte development for brain infiltration"
+Repository with code to reproduce analyses and figures in "Neurodevelopmental hijacking of oligodendrocyte lineage programs drives glioblastoma infiltration"
 
 
 ## Abstract
 
-Glioblastoma (GBM) is a highly aggressive brain tumor known for its ability to infiltrate and colonize normal brain tissue. Despite the clinical relevance of this behavior, the molecular characteristics of infiltrating GBM cells in the peritumoral zone remain largely unexplored in patients. Here, we show that peritumoral progenitor-like GBM cells overexpress transcriptional programs associated with increased neuronal migration, synaptic activity, and NOTCH signaling. Moreover, they spatially colocalize with neurons and have increased propensity for neuronal crosstalk. These infiltrative cells hijack key transcription factors that typically direct normal oligodendrocyte progenitor cell (OPC) differentiation. We further demonstrate that the epigenetic encoding of these infiltrative cells mirrors that of uncommitted OPCs in the developing human brain, a neurodevelopmental state marked by increased synaptic and migratory capabilities. Our findings provide critical insights into the neurodevelopmental hijacking that drives GBM infiltration in patients, rationalizing further investigation into targeting differentiation potential as a therapeutic strategy. 
+Glioblastoma (GBM) is a highly aggressive brain tumor known for its ability to infiltrate and colonize normal brain tissue. Despite the clinical relevance of this behavior, the molecular underpinnings of infiltrating GBM cells in the peritumoral zone remain unexplored in patients. Here, we show that peritumoral progenitor-like GBM cells activate transcriptional programs associated with increased invasivity, synaptic activity, and NOTCH signaling. These cells spatially colocalize with neurons and exhibit increased propensity for neuronal crosstalk. The epigenetic encoding of these infiltrative cells mirrors that of uncommitted OPCs in the developing human brain, a neurodevelopmental state marked by increased synaptic and migratory potential. Functional perturbation of a nominated regulatory factor, ZEB1, confirmed its role in maintaining the invasive and uncommitted developmental potential of infiltrative GBM cells. Our findings provide insights into the neurodevelopmental hijacking that drives GBM infiltration in patients, rationalizing further investigation into targeting differentiation potential as a therapeutic strategy.
+
+## Analysis workflow
+![Analysis Workflow](misc/Workflow.png)
 
 ## Contact
 

air.toml

@@ -0,0 +1,8 @@
+[format]
+line-width = 80
+indent-width = 4
+indent-style = "space"
+line-ending = "auto"
+persistent-line-breaks = true
+exclude = [".nextflow/", "slurm_out", "Python"]
+default-exclude = true

analysis/Fig1a.R

@@ -0,0 +1,152 @@
+# ---- Code to reproduce Figure 1a ---- #
+
+# Unload all previously loaded packages + remove previous environment
+rm(list = ls(all = TRUE))
+pacman::p_unload()
+
+# Set working directory
+GaitiLabUtils::set_wd()
+
+# ---- Setup script ---- #
+
+# Load required packages
+pacman::p_load(dplyr, ComplexHeatmap, RColorBrewer, tidyr, circlize)
+
+# Helper function for drawing Heatmaps
+cell_border_fun <- function(j, i, x, y, width, height, fill) {
+    grid.rect(
+        x = x,
+        y = y,
+        width = width,
+        height = height,
+        gp = gpar(col = "black", lwd = 1, fill = NA)
+    )
+}
+
+# Required inputs
+params <- list(
+    plot_dir = "output/figures",
+    input_table_path = "misc/Table S1.xlsx" # can be downloaded online, see publication
+)
+
+GaitiLabUtils::create_dir(params$plot_dir)
+
+# ---- Load data & Data wrangling ---- #
+mutations_oi <- c("IDH", "CDKN2A", "TP53", "ATRX", "BRAF")
+regions_oi <- c("PT", "TE", "TC")
+
+df <- readxl::read_excel(
+    params$input_table_path,
+    sheet = "Patient and sample info",
+    skip = 2
+) %>%
+    tidyr::fill(everything()) %>%
+    mutate(
+        IDH = ifelse(
+            stringr::str_detect(`Molecular alterations`, "IDH wildtype"),
+            "WT",
+            "MUT"
+        ),
+        CDKN2A = ifelse(
+            stringr::str_detect(
+                `Molecular alterations`,
+                "CDKN2A homozygous deletion"
+            ),
+            "DEL",
+            "WT"
+        ),
+        TP53 = ifelse(
+            stringr::str_detect(
+                `Molecular alterations`,
+                "TP53 p\\.C277Wfs\\*27  mutation"
+            ) |
+                stringr::str_detect(`Molecular alterations`, "P53 Mutated"),
+            "MUT",
+            "WT"
+        ),
+        ATRX = ifelse(
+            stringr::str_detect(`Molecular alterations`, "ATRX retained"),
+            "WT",
+            "MUT"
+        ),
+        BRAF = ifelse(
+            stringr::str_detect(`Molecular alterations`, "BRAF mutation"),
+            "MUT",
+            "WT"
+        ),
+    ) %>%
+    group_by_at(c(
+        "Patient ID",
+        "Sex",
+        "IDH",
+        "CDKN2A",
+        "TP53",
+        "ATRX",
+        "BRAF",
+        "Region annotation"
+    )) %>%
+    summarise(n = n()) %>%
+    pivot_wider(names_from = "Region annotation", values_from = "n")
+
+col_annot_df <- df %>%
+    ungroup() %>%
+    dplyr::select(all_of(c("Sex"))) %>%
+    pull()
+
+
+# ---- Create Figure 1a ---- #
+
+col_annot <- HeatmapAnnotation(
+    Sex = col_annot_df,
+    which = "col",
+    col = list(
+        Sex = c("Male" = "white", "Female" = "black")
+    ),
+    gp = gpar(col = "black")
+)
+
+heatmap_df <- df %>%
+    ungroup() %>%
+    dplyr::select(all_of(mutations_oi))
+heatmap_df <- as.matrix(heatmap_df)
+rownames(heatmap_df) <- df %>% pull(`Patient ID`)
+ht1 <- Heatmap(
+    t(heatmap_df),
+    name = "Molecular features",
+    col = c("WT" = "white", "MUT" = "grey", "DEL" = "black"),
+    cluster_rows = FALSE,
+    cluster_columns = FALSE,
+    border = TRUE,
+    cell_fun = cell_border_fun,
+    show_column_names = FALSE,
+    top_annotation = col_annot,
+)
+
+heatmap_df2 <- df %>%
+    ungroup() %>%
+    dplyr::select(all_of(regions_oi))
+heatmap_df2 <- as.matrix(heatmap_df2)
+rownames(heatmap_df2) <- df %>% pull(`Patient ID`)
+heatmap_df2 <- ifelse(is.na(heatmap_df2), 0, heatmap_df2)
+col_scheme <- brewer.pal(5, "Blues")
+col_scheme[1] <- "white"
+
+ht2 <- Heatmap(
+    t(heatmap_df2),
+    name = "Number of samples from region",
+    col = circlize::colorRamp2(c(0, 1, 2, 3, 4), col_scheme),
+    cluster_rows = FALSE,
+    cluster_columns = FALSE,
+    border = TRUE,
+    cell_fun = cell_border_fun
+)
+
+heatmap <- ht1 %v% ht2
+
+pdf(
+    file = file.path(params$plot_dir, "Fig1a.pdf"),
+    width = 10,
+    height = 10
+)
+draw(heatmap, ht_gap = unit(0.5, "cm"))
+dev.off()