Functional genomic profiling of schizophrenia-associated genes reveals key microglial regulators.
4 blocks of code here have a permanent identifier (oscr:b:…), listed with the matches below: citable for good, with a page of its own. Select a match and press Cite to copy it. What a block identifier is.
The 7 matches
- [1] § Results › Transcriptomic profiling reveals diverse gene targeting effects on microglial state ↔ image_analysis/Crispr30_publication_final.Rmd, lines 393 to 469 · score 0.86 · HLA DMB, SLC11A1, FCGR3B, NCKAP1L, ARHGAP25, BIN2 · block
oscr:b:VH0FN9N1PXR7T-5· Report a mismatch - [2] § Results › Transcriptomic profiling reveals diverse gene targeting effects on microglial state ↔ drug_seq/generate_string_plot_final.Rmd, lines 32 to 48 · score 0.77 · HLA DMB, SLC11A1, NCKAP1L, ARHGAP25, BIN2, TLR6 · block
oscr:b:QR0ZX73M18JSV-Y· Report a mismatch - [3] § Results › Functional programs induced by targeted genes ↔ drug_seq/03_01_analyze_deg_go.R, lines 441 to 500 · score 0.66 · cell adhesion, complement activation, protein production, cytokine, GO, chemokine · block
oscr:b:THPNK0W85WHYX-R· Report a mismatch - [4] § Results › Functional programs induced by targeted genes ↔ image_analysis/Crispr30_publication_final.Rmd, lines 393 to 469 · score 0.59 · CLEC7A, FCGR2A, CD14, IRF8, SYK, TREM2 · block
oscr:b:VH0FN9N1PXR7T-5· Report a mismatch - [5] § Results › Functional programs induced by targeted genes ↔ drug_seq/generate_string_plot_final.Rmd, lines 32 to 48 · score 0.58 · CLEC7A, FCGR2A, CD14, IRF8, SYK, TREM2 · block
oscr:b:QR0ZX73M18JSV-Y· Report a mismatch - [6] § Results › Transcriptomic profiling reveals diverse gene targeting effects on microglial state ↔ drug_seq/03_01_analyze_deg_go.R, lines 113 to 229 · score 0.52 · log2 fold change, log2FC, Asterisks, Pairwise, pvalue, DEGs · block
oscr:b:588RGJHKYNCGX-G· Report a mismatch - [7] § Results › Functional programs induced by targeted genes ↔ drug_seq/03_01_analyze_deg_go.R, lines 113 to 229 · score 0.51 · log2 fold change, log2FC, Asterisks, GO, heatmaps, DEG · block
oscr:b:588RGJHKYNCGX-G· Report a mismatch
Paper
Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC
The paper is loaded when this pane is shown.
The authors' code
R · 1,454 lines · 49 KB · no license · 3 matches
- LIB='/cluster/tufts/patralab/rbator01/R_libs/4.4.0'
- .libPaths(c("",LIB))
- library(pheatmap)
- library(tidyverse)
- library(ComplexHeatmap)
- library(openxlsx)
- library(ggforce)
- library(data.table)
- library(clusterProfiler)
- library('org.Hs.eg.db')
- library(data.table)
- library(circlize)
- library(grid)
- library(forcats)
- library(stringr)
- library(janitor)
- library(DESeq2)
- library(ggbreak)
- select = dplyr::select
- rename = dplyr::rename
- filter=dplyr::filter
- # paths and setup ----
- setwd('/cluster/tufts/patralab/rbator01/perlis_lab/cspr_apr25/')
- # DEG thresholds
- padj_thresh = 0.1
- praw_thresh = 0.1
- # load metadata and counts ----
- meta = read.xlsx("analysis/metadata_batch1/241210_Joy30_Metadata_format.xlsx")
- ctrl_counts = read.xlsx("analysis/nct1_expression/nct1_cpm_ko_genes.xlsx")
- rm_samples = fread("analysis/de/removed_samples_control_batch_q1q3_filtered_batch1_batch2.csv") %>%
- mutate(tmp = name) %>%
- separate(tmp, into=c("ko_gene","well","batch"), sep="\\|")
- n_rm = rm_samples %>%
- filter(batch == "b1") %>%
- group_by(ko_gene) %>%
- mutate(n_rm = n()) %>%
- ungroup() %>%
- as.data.frame()
- # deg results b1 -----
- b1 = fread("analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv")
- head(b1)
- res_sig_b1 = b1 %>%
- filter(pvalue < praw_thresh)
- # ko_gene in gene_name at praw < praw_thresh
- ko_in_praw_df_b1 <- res_sig_b1 %>%
- group_by(ko_gene) %>%
- summarise(ko_in_praw_deg_b1 = as.integer(ko_gene %in% gene_name), .groups = "drop")%>%
- distinct()
- # ko_gene in gene_name at padj < padj_thresh
- ko_in_padj_df_b1 <- res_sig_b1 %>%
- filter(padj < padj_thresh) %>%
- group_by(ko_gene) %>%
- summarise(ko_in_padj_deg_b1 = as.integer(ko_gene %in% gene_name), .groups = "drop")%>%
- distinct()
- # count DEGs
- t_b1 = as.data.frame(table((b1 %>% filter(padj < padj_thresh))$ko_gene))
- colnames(t_b1) = c("ko_gene","ndeg_padj_0.1_b1")
- top_b1 <- b1 %>%
- group_by(ko_gene) %>%
- arrange(padj) %>%
- slice_min(order_by = padj, n = 10, with_ties = FALSE) %>%
- select(ko_gene, gene_name) %>%
- rename(top_b1 = gene_name)
- top_b1_collapsed <- aggregate(top_b1 ~ ko_gene, data = top_b1, FUN = function(x) paste(x, collapse = ","))
- # assemble summary workbook ---------
- final = ctrl_counts %>%
- rename(ko_gene=gene_name) %>%
- mutate(coverage_pass = ifelse(mean_cpm > 5, 1,0)) %>%
- full_join(n_rm %>% dplyr::select(ko_gene, n_rm), by="ko_gene") %>%
- left_join(t_b1, by = "ko_gene") %>%
- left_join(ko_in_padj_df_b1, by = "ko_gene") %>%
- left_join(ko_in_praw_df_b1, by = "ko_gene") %>%
- left_join(top_b1_collapsed, by = "ko_gene")
- # Replace NA with 0
- final <- final %>%
- mutate(across(starts_with("ko_in_padj_deg_"), ~ if_else(is.na(.), 0L, .))) %>%
- mutate(across(starts_with("ko_in_praw_deg_"), ~ if_else(is.na(.), 0L, .))) %>%
- mutate(across(starts_with("ndeg_padj_0.1_"), ~ if_else(is.na(.), 0L, as.integer(.))))
- final = final %>%
- distinct()
- write.xlsx(final, "analysis/de/summary_batch1_q1q3_filter_rmlow_22aug25.xlsx")
- # Omics Fig 1A: DEG target overlap, batch 1 only =======
- dds = readRDS("analysis/processed/umi.trim.1mm_all_dedup.counts.dds.rds")
- genes_rm = c('RBFOX3','FCGR3B','PLD4','FCAR','NTC1')
- meta = colData(dds) %>%
- as.data.frame() %>%
- filter(!(KO.Target %in% genes_rm))
- all_targets = unique(meta$KO.Target)
- all_targets = factor(all_targets, levels = all_targets)
- ## Helper: build KO–KO log2FC heatmap with asterisks for significant entries
- make_heat <- function(deg_file,
- title_text,
- out_pdf,
- lfc_type = c("raw", "shrunken"),
- sig_col = c("padj", "pvalue"),
- sig_thresh = 0.1) {
- lfc_type <- match.arg(lfc_type) # "raw" or "shrunken"
- sig_col <- match.arg(sig_col) # "padj" or "pvalue"
- # Append info to title for clarity
- #title_text <- paste0(title_text, " ", lfc_type, " (", sig_col, " < ", sig_thresh, ")")
- # read & standardize column names
- deg <- data.table::fread(deg_file) %>%
- dplyr::rename(gene = gene_name)
- # ensure chosen significance column is present
- if (!sig_col %in% colnames(deg)) {
- stop(sprintf("Column '%s' not found in '%s'.", sig_col, deg_file))
- }
- # choose LFC column
- lfc_col <- if (lfc_type == "raw") "log2FoldChange_raw" else "log2FoldChange"
- # keep targets, aggregate duplicates:
- # - LFC columns: mean
- # - star_flag: TRUE if any chosen sig_col < sig_thresh
- agg <- deg %>%
- dplyr::select(gene, ko_gene, log2FoldChange_raw, log2FoldChange, pvalue, padj) %>%
- dplyr::filter(gene %in% all_targets, ko_gene %in% all_targets) %>%
- dplyr::group_by(gene, ko_gene) %>%
- dplyr::summarise(
- log2FoldChange_raw = mean(log2FoldChange_raw, na.rm = TRUE),
- log2FoldChange = mean(log2FoldChange, na.rm = TRUE),
- star_flag = any(.data[[sig_col]] < sig_thresh, na.rm = TRUE),
- .groups = "drop"
- ) %>%
- dplyr::mutate(dplyr::across(c(log2FoldChange_raw, log2FoldChange),
- ~ ifelse(is.nan(.x), NA_real_, .x)))
- # complete to full target grid; impute chosen LFC to 0
- heat_long <- agg %>%
- tidyr::complete(
- gene = all_targets,
- ko_gene = all_targets
- ) %>%
- dplyr::mutate(
- value = .data[[lfc_col]],
- value = tidyr::replace_na(value, 0),
- star_flag = tidyr::replace_na(star_flag, FALSE),
- gene = factor(gene, levels = all_targets),
- ko_gene = factor(ko_gene, levels = all_targets)
- )
- # value matrix
- heat_mat <- heat_long %>%
- dplyr::select(gene, ko_gene, value) %>%
- tidyr::pivot_wider(names_from = ko_gene, values_from = value) %>%
- tibble::column_to_rownames("gene") %>%
- as.matrix()
- # asterisk (logical) matrix
- star_mat <- heat_long %>%
- dplyr::select(gene, ko_gene, star_flag) %>%
- tidyr::pivot_wider(names_from = ko_gene, values_from = star_flag) %>%
- tibble::column_to_rownames("gene") %>%
- as.matrix()
- ## cluster columns (correlation distance)
- d_col <- as.dist(1 - stats::cor(heat_mat, use = "pairwise.complete.obs"))
- col_clust <- stats::hclust(d_col, method = "average")
- col_order_names <- colnames(heat_mat)[col_clust$order]
- # row order aligned to clustered columns (append non-overlapping rows)
- common <- intersect(col_order_names, rownames(heat_mat))
- first_part <- match(common, rownames(heat_mat))
- row_order <- c(first_part, setdiff(seq_len(nrow(heat_mat)), first_part))
- row_gp <- grid::gpar(fontfamily = "Arial", fontsize = 8)
- col_gp <- grid::gpar(fontfamily = "Arial", fontsize = 8)
- title_gp <- grid::gpar(fontfamily = "Arial", fontsize = 12, fontface = "bold")
- legend_gp <- grid::gpar(fontfamily = "Arial")
- ht <- ComplexHeatmap::Heatmap(
- heat_mat,
- name = "log2FC", #ifelse(lfc_type == "raw", "log2FC_raw", "log2FC_shrunken"),
- col = circlize::colorRamp2(
- c(min(heat_mat, na.rm = TRUE), 0, max(heat_mat, na.rm = TRUE)),
- c("blue", "white", "red")
- ),
- cluster_columns = as.dendrogram(col_clust),
- cluster_rows = FALSE,
- row_order = row_order,
- show_row_names = TRUE,
- show_column_names = TRUE,
- row_names_gp = row_gp,
- column_names_gp = col_gp,
- column_title_gp = title_gp,
- heatmap_legend_param = list(
- labels_gp = legend_gp,
- title_gp = grid::gpar(fontfamily = "Arial", fontface = "bold")
- ),
- cell_fun = function(j, i, x, y, width, height, fill) {
- if (!is.na(star_mat[i, j]) && star_mat[i, j]) {
- grid::grid.text("*", x = x, y = y,
- gp = grid::gpar(fontfamily = "Arial", fontsize = 10, fontface = "bold"))
- }
- }
- )
- # Use Cairo to embed Arial cleanly in the PDF
- Cairo::CairoPDF(out_pdf, height = 5, width = 6, family = "Arial")
- on.exit(dev.off(), add = TRUE)
- ComplexHeatmap::draw(ht)
- }
- # KO–KO overlap using raw LFC and p-value < 0.1
- make_heat(
- deg_file = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- title_text = "",
- out_pdf = "analysis/de/plots/deg_target_overlap_b1_lfcraw_pval_0.1.pdf",
- lfc_type = "raw",
- sig_col = "pvalue",
- sig_thresh = 0.1 )
- make_heat(
- deg_file = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- title_text = "",
- out_pdf = "analysis/de/plots/deg_target_overlap_b1_lfcraw_padj_0.1.pdf",
- lfc_type = "raw",
- sig_col = "padj",
- sig_thresh = 0.1 )
- # Barplot NUMBER DEG =======
- tab <- b1 %>%
- group_by(ko_gene) %>%
- summarise(n = sum(padj < 0.1, na.rm = TRUE), .groups = "drop")
- # Color/category per KO from the *diagonal* row (ko_gene == gene_name)
- diag_map <- b1 %>%
- filter(ko_gene == gene_name) %>%
- transmute(
- ko_gene,
- p_raw = pvalue,
- lfc = dplyr::coalesce(log2FoldChange, log2FoldChange_raw),
- direction = case_when(
- is.finite(p_raw) & p_raw < 0.1 & is.finite(lfc) & lfc > 0 ~ "sig_up",
- is.finite(p_raw) & p_raw < 0.1 & is.finite(lfc) & lfc < 0 ~ "sig_down",
- TRUE ~ "other"
- )
- )
- tab <- tab %>%
- left_join(diag_map[, c("ko_gene", "direction")], by = "ko_gene") %>%
- mutate(direction = tidyr::replace_na(direction, "other"),
- direction = factor(direction, levels = c("sig_up", "sig_down", "other")))
- # Find outlier and compute a sensible break
- top_row <- tab %>% slice_max(n, n = 1, with_ties = FALSE)
- top_val <- top_row$n
- max_others <- max(tab$n[tab$ko_gene != top_row$ko_gene], na.rm = TRUE)
- gap <- max(5, floor(0.05 * (top_val - max_others))) # ~5% gap or ≥5
- lower_brk <- max_others + gap
- upper_brk <- top_val - gap
- # Axis ticks: show normal ticks below break + a single tick at 1100 above break
- lower_ticks <- pretty(c(0, max_others))
- upper_tick <- 1100
- all_breaks <- unique(c(lower_ticks, upper_tick))
- p <- ggplot(tab, aes(x = fct_reorder(ko_gene, n), y = n, fill = direction)) +
- geom_col(color = "black") +
- coord_flip(clip = "off") +
- labs(
- x = "Target Gene",
- y = "n (DEGs with padj < 0.1)",
- title = "",
- fill = "DEG status of Target Gene"
- ) +
- scale_fill_manual(
- values = c(sig_up = "red", sig_down = "blue", other = "black"),
- labels = c(
- sig_up = "p < 0.1 & log2FC > 0",
- sig_down = "p < 0.1 & log2FC < 0",
- other = "otherwise"
- )
- ) +
- # Broken axis
- ggbreak::scale_y_break(c(lower_brk, upper_brk)) +
- # Only label lower ticks + one tick (1100) in the upper segment
- scale_y_continuous(
- breaks = all_breaks,
- labels = function(b) ifelse(b == upper_tick, "1100", b)
- ) +
- theme_minimal(base_size = 12, base_family = "Arial") +
- theme(
- text = element_text(family = "Arial"),
- plot.title = element_text(face = "bold"),
- legend.position = "right",
- plot.margin = margin(5.5, 20, 5.5, 5.5)
- ) +
- # ensure the 1100 tick is in range even if top_val < 1100
- expand_limits(y = max(top_val, upper_tick) * 1.02)
- systemfonts::match_font("Arial")
- show(p)
- ggsave("analysis/de/plots/deg_per_ko_gene.pdf", p,width = 7.5, height = 5, device = cairo_pdf, family = "Arial",onefile = FALSE)
- # Same barplot but all blue (no categories in legend)
- p <- ggplot(tab, aes(x = fct_reorder(ko_gene, n), y = n)) +
- geom_col(color = "black", fill = "blue") +
- coord_flip(clip = "off") +
- labs(
- x = "Target Gene",
- y = "n (DEGs with padj < 0.1)",
- title = ""
- ) +
- ggbreak::scale_y_break(c(lower_brk, upper_brk)) +
- scale_y_continuous(
- breaks = all_breaks,
- labels = function(b) ifelse(b == upper_tick, "1100", b)
- ) +
- theme_minimal(base_size = 12, base_family = "Arial") +
- theme(
- text = element_text(family = "Arial"),
- plot.title = element_text(face = "bold"),
- legend.position = "none", # hide legend
- plot.margin = margin(5.5, 20, 5.5, 5.5)
- ) +
- expand_limits(y = max(top_val, upper_tick) * 1.02)
- print(p)
- ggsave("analysis/de/plots/deg_per_ko_gene_blue.pdf", p,
- width = 5, height = 5, device = cairo_pdf, family = "Arial", onefile = FALSE)
- # GO ENRICHMENT for targets that have >10 DEG =================
- b1_filter = fread("analysis/de/all_res_vs_ntc1_padj_0.1_10deg_batch1_q1q3_filter_rmlow.tsv")
- ck<- compareCluster(geneCluster = ens ~ ko_gene,
- data = b1_filter,
- OrgDb = org.Hs.eg.db,
- keyType="ENSEMBL",
- fun = "enrichGO",
- ont="BP",
- readable=T)
- saveRDS(ck, "analysis/de/go_bp_batch1_padj_0.1_10deg.rds")
- write.xlsx(ck, "analysis/de/go_bp_batch1_padj_0.1_10deg.xlsx")
- ck<- compareCluster(geneCluster = ens ~ ko_gene,
- data = b1_filter,
- OrgDb = org.Hs.eg.db,
- keyType="ENSEMBL",
- fun = "enrichGO",
- ont="CC",
- readable=T)
- saveRDS(ck, "analysis/de/go_cc_batch1_padj_0.1_10deg.rds")
- write.xlsx(ck, "analysis/de/go_cc_batch1_padj_0.1_10deg.xlsx")
- ck<- compareCluster(geneCluster = ens ~ ko_gene,
- data = b1_filter,
- OrgDb = org.Hs.eg.db,
- keyType="ENSEMBL",
- fun = "enrichGO",
- ont="MF",
- readable=T)
- saveRDS(ck, "analysis/de/go_mf_batch1_padj_0.1_10deg.rds")
- write.xlsx(ck, "analysis/de/go_mf_batch1_padj_0.1_10deg.xlsx")
- ## Quick sanity plots for GO enrichments (BP / MF / CC)
- ck = readRDS("analysis/de/go_bp_batch1_padj_0.1.rds")
- dotplot(ck, show=3)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- ck = readRDS("analysis/de/go_mf_batch1_padj_0.1.rds")
- dotplot(ck, show=3)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- ck = readRDS("analysis/de/go_cc_batch1_padj_0.1.rds")
- dotplot(ck, show=3)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- ## GO BP: filter by keyword families ----
- keywords <- c(
- "phago", "autophag",
- "neuro",
- "macrophage", "activation",
- "adhesion", "attachment", "integrin", "actin",
- "chemotaxis", "chemokine", "cytokine",
- "inflamm", "stress",
- "lyso",
- "metabolic", "oxydat",
- "mitochondria"
- )
- ck = readRDS("analysis/de/go_bp_batch1_padj_0.1_10deg.rds")
- # Collapse keywords into regex pattern
- pattern <- paste(keywords, collapse = "|")
- # Filter the dataframe based on keyword match in the specified column
- ck_filter = ck %>%
- filter(grepl(pattern, ck@compareClusterResult$Description, ignore.case = TRUE))
- dotplot(ck_filter, show=3)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- # Select terms from collaborators for initial combination -----
- #BP
- ck = readRDS("analysis/de/go_bp_batch1_padj_0.1_10deg.rds")
- s = read.xlsx("analysis/de/go_bp_batch1_padj_0.1_10deg_sds.xlsx") %>%
- mutate(select = ifelse(select == 1, 's',select)) %>%
- filter(select == 's') %>%
- select(ID, select)
- l = read.xlsx("analysis/de/go_bp_batch1_padj_0.1_10deg_LM.xlsx") %>%
- filter(!is.na(LM)) %>%
- rename(select = LM) %>%
- mutate(select = 'l') %>%
- select(ID, select)
- c = read.xlsx("analysis/de/go_bp_batch1_padj_0.1_10deg.xlsx") %>%
- filter(!is.na(CAMILLA_list)) %>%
- rename(select = CAMILLA_list ) %>%
- select(ID, select)
- all_select = rbind(s,l,c) %>%
- distinct()
- dedup_select <- all_select %>%
- mutate(
- is_num = grepl("^\\d+$", select),
- # convert only the numeric entries; letters -> NA (no warning)
- num_val = as.integer(replace(select, !is_num, NA_character_)),
- letter_score = case_when(
- !is_num & select == "s" ~ 0L, # "s" beats other letters
- !is_num ~ 1L,
- TRUE ~ NA_integer_
- ),
- key1 = if_else(is_num, 0L, 1L), # numbers (0) beat letters (1)
- key2 = if_else(is_num, num_val, letter_score) # among numbers: smallest wins; among letters: "s" wins
- ) %>%
- arrange(ID, key1, key2) %>%
- distinct(ID, .keep_all = TRUE) %>% # one row per GO ID
- select(ID, select)
- ck_select = ck@compareClusterResult %>%
- left_join(dedup_select, by="ID") %>%
- filter(p.adjust < 0.05)
- lab_map <- c(
- `1` = "Phagocytosis / Complement activation",
- `2` = "Cell adhesion",
- `3` = "Chemotaxis",
- `4` = "PH regulation",
- `5` = "Chemokines/Cytokines",
- `6` = "Protein production",
- l = "liam",
- s = "steve"
- )
- ck_select$select_label <- lab_map[as.character(ck_select$select)]
- write.xlsx(ck_select, "analysis/de/go_bp_batch1_padj_0.1_10deg_allselect.xlsx")
- # MF ------
- ck = readRDS("analysis/de/go_mf_batch1_padj_0.1")
- l = read.xlsx("analysis/de/go_mf_batch1_padj_0.1_10deg_LM.xlsx") %>%
- filter(!is.na(LM)) %>%
- rename(select = LM) %>%
- mutate(select = 'l') %>%
- select(ID, select)
- c = read.xlsx("analysis/de/go_mf_batch1_padj_0.1_10deg.xlsx") %>%
- filter(!is.na(Camilla_list)) %>%
- rename(select = Camilla_list ) %>%
- select(ID,select)
- all_select = rbind(l,c) %>%
- distinct()
- dedup_select <- all_select %>%
- mutate(
- is_num = grepl("^\\d+$", select),
- # convert only the numeric entries; letters -> NA (no warning)
- num_val = as.integer(replace(select, !is_num, NA_character_)),
- letter_score = case_when(
- !is_num & select == "s" ~ 0L, # "s" beats other letters
- !is_num ~ 1L,
- TRUE ~ NA_integer_
- ),
- key1 = if_else(is_num, 0L, 1L), # numbers (0) beat letters (1)
- key2 = if_else(is_num, num_val, letter_score) # among numbers: smallest wins; among letters: "s" wins
- ) %>%
- arrange(ID, key1, key2) %>%
- distinct(ID, .keep_all = TRUE) %>% # one row per GO ID
- select(ID, select)
- ck_select = ck %>%
- filter(ID %in% dedup_select$ID)
- ck_select_sim <- simplify(ck_select, cutoff = 0.7, by = "p.adjust", select_fun = min)
- dotplot(ck_select_sim, show=200)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- ck_select_sim_df = ck_select_sim@compareClusterResult %>%
- left_join(dedup_select, by="ID")
- lab_map <- c(
- `1` = "Phagocytosis / Complement activation",
- `2` = "Cell adhesion",
- `3` = "Chemotaxis",
- `4` = "PH regulation",
- `5` = "Chemokines/Cytokines",
- `6` = "Protein production",
- l = "liam",
- s = "steve"
- )
- ck_select_sim_df$select_label <- lab_map[as.character(ck_select_sim_df$select)]
- write.xlsx(ck_select_sim_df, "analysis/de/go_mf_batch1_padj_0.1_10deg_allselect.xlsx")
- # CC -----
- ck = readRDS("analysis/de/go_cc_batch1_padj_0.1_10deg.rds")
- l = read.xlsx("analysis/de/go_cc_batch1_padj_0.1_10deg_LM.xlsx") %>%
- filter(!is.na(LM)) %>%
- rename(select = LM) %>%
- mutate(select = 'l') %>%
- select(ID, select)
- c = read.xlsx("analysis/de/go_cc_batch1_padj_0.1_10deg.xlsx") %>%
- filter(!is.na(Camilla_list)) %>%
- rename(select = Camilla_list ) %>%
- select(ID,select)
- all_select = rbind(l,c) %>%
- distinct()
- dedup_select <- all_select %>%
- mutate(
- is_num = grepl("^\\d+$", select),
- # convert only the numeric entries; letters -> NA (no warning)
- num_val = as.integer(replace(select, !is_num, NA_character_)),
- letter_score = case_when(
- !is_num & select == "s" ~ 0L, # "s" beats other letters
- !is_num ~ 1L,
- TRUE ~ NA_integer_
- ),
- key1 = if_else(is_num, 0L, 1L), # numbers (0) beat letters (1)
- key2 = if_else(is_num, num_val, letter_score) # among numbers: smallest wins; among letters: "s" wins
- ) %>%
- arrange(ID, key1, key2) %>%
- distinct(ID, .keep_all = TRUE) %>% # one row per GO ID
- select(ID, select)
- ck_select = ck %>%
- filter(ID %in% dedup_select$ID)
- ck_select_sim <- simplify(ck_select, cutoff = 0.7, by = "p.adjust", select_fun = min)
- dotplot(ck_select_sim, show=200)+
- theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))+
- scale_y_discrete(labels = ~ str_wrap(as.character(.x), 100))
- ck_select_sim_df = ck_select_sim@compareClusterResult %>%
- left_join(dedup_select, by="ID")
- lab_map <- c(
- `1` = "Phagocytosis / Complement activation",
- `2` = "Cell adhesion",
- `3` = "Chemotaxis",
- `4` = "PH regulation",
- `5` = "Chemokines/Cytokines",
- `6` = "Protein production",
- l = "liam",
- s = "steve"
- )
- ck_select_sim_df$select_label <- lab_map[as.character(ck_select_sim_df$select)]
- write.xlsx(ck_select_sim_df, "analysis/de/go_cc_batch1_padj_0.1_10deg_allselect.xlsx")
- # make a plot of the CC individual results with collaborator selections
- # ---------- helpers ----------
- safe_min <- function(x) if (all(is.na(x))) NA_real_ else min(x, na.rm = TRUE)
- strip_go_brackets <- function(x) sub("\\s*\\[[^\\]]+\\]$", "", x)
- # choose one 'select' per y_id:
- # - if any numbers → pick the smallest number
- # - else letters only → prefer "s", otherwise first other letter
- choose_select <- function(x) {
- xo <- as.character(x)
- nums <- suppressWarnings(as.integer(xo))
- if (any(!is.na(nums))) return(as.character(min(nums, na.rm = TRUE)))
- if (any(xo == "s")) return("s")
- setdiff(unique(xo), "s")[1]
- }
- # Start from ck_select_sim_df; map columns ----------
- base_df <- ck_select_sim_df %>%
- transmute(
- comparison = Cluster,
- GO_ID = ID,
- GO_Term = Description,
- padj = p.adjust,
- Significant_genes_count = Count,
- select
- )
- # OPTIONAL: union of Top-10 terms per comparison by padj
- top_ids <- base_df %>%
- group_by(comparison) %>%
- slice_min(order_by = padj, n = 100, with_ties = FALSE) %>%
- pull(GO_ID) %>% unique()
- base_df <- base_df %>% filter(GO_ID %in% top_ids)
- # Build plot_df ----------
- eps <- 1e-300
- cap <- 10
- # y_id is now JUST the term
- plot_df <- base_df %>%
- mutate(
- mlog10 = pmin(-log10(pmax(padj, eps)), cap),
- y_id = GO_Term
- ) %>%
- group_by(y_id, comparison) %>%
- summarise(
- mlog10 = if (all(is.na(mlog10))) NA_real_ else max(mlog10, na.rm = TRUE),
- Significant_genes_count = if (all(is.na(Significant_genes_count)))
- NA_real_ else max(Significant_genes_count, na.rm = TRUE),
- .groups = "drop"
- ) %>%
- # annotation: collapse select per DESCRIPTION
- left_join(
- base_df %>%
- transmute(y_id = GO_Term, select) %>%
- group_by(y_id) %>%
- summarise(select = choose_select(select), .groups = "drop"),
- by = "y_id"
- )
- # order y by best padj across comparisons
- y_levels <- base_df %>%
- group_by(GO_Term) %>%
- summarise(best_padj = safe_min(padj), .groups = "drop") %>%
- arrange(best_padj) %>%
- pull(GO_Term) %>% unique()
- plot_df <- plot_df %>% mutate(y_id = factor(y_id, levels = y_levels))
- # DOTPLOT of −log10(padj) ----------
- gg_dot <- ggplot(plot_df, aes(x = comparison, y = y_id)) +
- geom_point(aes(size = Significant_genes_count, color = mlog10), na.rm = TRUE) +
- scale_size_continuous(name = "Significant genes", range = c(1.8, 8)) +
- scale_color_viridis_c(name = expression(-log[10](padj)), option = "C",
- direction = 1, na.value = "grey85") +
- labs(title = "Selected GO Terms",
- subtitle = sprintf("Colored by −log10(padj), capped at %s;\n size = Significant_genes_count", cap),
- x = "Comparison", y = NULL) +
- theme_bw() +
- theme(axis.text.y = element_text(size = 9),
- axis.text.x = element_text(angle = 45, hjust = 1))
- print(gg_dot)
- # FINAL plot with all annotations. =====
- # Target order from functional centered on NEG
- well_summary=read.xlsx('20250415_Joshua_Bowen_8511/2025822_functionaldata_celllevel_exp31_only_to_show_ipa_go.xlsx')
- exp31 = well_summary %>%
- filter(Experiment == "EXP31")
- control <- "NEG"
- average_by_experiment_treatment_well <- function(cell_data) {
- cell_data %>%
- group_by(Experiment, Treatment, Metadata_Well) %>%
- summarise(
- across(AreaShape_Area:Cell_Level_Phago_Index, mean, na.rm = TRUE),
- .groups = "drop"
- )
- }
- neg_mean <- well_summary %>%
- filter(Treatment == control) %>%
- summarise(m = mean(Cell_Level_Phago_Index, na.rm = TRUE), .groups = "drop") %>%
- pull(m)
- centered_order_tbl <- well_summary %>%
- group_by(Treatment) %>%
- summarise(mean_clpi = mean(Cell_Level_Phago_Index, na.rm = TRUE), .groups = "drop") %>%
- mutate(delta_clpi = mean_clpi - neg_mean) %>%
- filter(Treatment != control) %>%
- arrange(desc(delta_clpi))
- # read GO dataframe ------
- df_ann <- openxlsx::read.xlsx("analysis/de/go_bp_batch1_padj_0.1_10deg_allselect_rebeccaedit.xlsx")
- df <- df_ann %>%
- janitor::clean_names() %>%
- mutate(
- select_label = as.character(select_label),
- select_label = stringr::str_squish(select_label),
- select_label = dplyr::na_if(select_label, "") # treat "" as NA
- ) %>%
- filter(!is.na(select_label)) %>%
- mutate(
- padj_safe = pmax(p_adjust, 1e-300) # guard tiny zeros
- )
- df_plot0 <- df %>%
- transmute(
- Pathway = description,
- Target = ko_gene,
- Count = count,
- Group = select_label,
- colval = padj_safe
- )
- # Keep only targets present in well_summary order
- go_targets <- unique(df_plot0$Target)
- target_levels <- centered_order_tbl %>%
- filter(Treatment %in% go_targets) %>%
- pull(Treatment)
- df_plot0 <- df_plot0 %>%
- filter(Target %in% target_levels)
- # count filter
- df_plot0 <- df_plot0 %>% filter(Count > 2)
- # ----------------------------
- # Collapse "X" with response/regulation variants
- # Keep, per base X: most treatments, then strongest mean -log10(padj),
- # then preference: exact < response < cellular response < positive < negative < regulation
- # ----------------------------
- strip_variants_to_base <- function(s) {
- s %>%
- # response variants
- stringr::str_replace(regex("^cellular\\s+response\\s+to\\s+", ignore_case = TRUE), "") %>%
- stringr::str_replace(regex("^response\\s+to\\s+", ignore_case = TRUE), "") %>%
- # regulation variants
- stringr::str_replace(regex("^positive\\s+regulation\\s+of\\s+", ignore_case = TRUE), "") %>%
- stringr::str_replace(regex("^negative\\s+regulation\\s+of\\s+", ignore_case = TRUE), "") %>%
- stringr::str_replace(regex("^regulation\\s+of\\s+", ignore_case = TRUE), "") %>%
- stringr::str_squish()
- }
- variant_type <- function(s) {
- if (stringr::str_detect(s, regex("^cellular\\s+response\\s+to\\s+", TRUE))) return("resp_cellular")
- if (stringr::str_detect(s, regex("^response\\s+to\\s+", TRUE))) return("response")
- if (stringr::str_detect(s, regex("^positive\\s+regulation\\s+of\\s+", TRUE))) return("pos")
- if (stringr::str_detect(s, regex("^negative\\s+regulation\\s+of\\s+", TRUE))) return("neg")
- if (stringr::str_detect(s, regex("^regulation\\s+of\\s+", TRUE))) return("reg")
- return("exact")
- }
- df_plot0 <- df_plot0 %>%
- mutate(
- base_name = strip_variants_to_base(Pathway),
- var_type = vapply(Pathway, variant_type, character(1))
- )
- # metrics per GO term across all targets
- path_metrics <- df_plot0 %>%
- group_by(Pathway, base_name) %>%
- summarise(
- n_treat = n_distinct(Target),
- mean_sig = mean(colval, na.rm = TRUE), # smaller is more significant
- .groups = "drop"
- )
- # preference tier: exact (1) < response (2) < cellular response (3) < positive (4) < negative (5) < regulation (6)
- path_metrics <- path_metrics %>%
- mutate(tier = dplyr::case_when(
- grepl("^cellular\\s+response\\s+to\\s+", Pathway, ignore.case = TRUE) ~ 3L,
- grepl("^response\\s+to\\s+", Pathway, ignore.case = TRUE) ~ 2L,
- grepl("^positive\\s+regulation\\s+of\\s+", Pathway, ignore.case = TRUE) ~ 4L,
- grepl("^negative\\s+regulation\\s+of\\s+", Pathway, ignore.case = TRUE) ~ 5L,
- grepl("^regulation\\s+of\\s+", Pathway, ignore.case = TRUE) ~ 6L,
- TRUE ~ 1L
- ))
- winners <- path_metrics %>%
- arrange(base_name, desc(n_treat), mean_sig, tier, Pathway) %>%
- group_by(base_name) %>%
- dplyr::slice(1) %>%
- ungroup()
- # keep only the winning GO term for each base X
- df_plot0 <- df_plot0 %>%
- semi_join(winners %>% dplyr::select(Pathway), by = "Pathway") %>%
- mutate(Group = factor(Group, levels = unique(Group)))
- # Keep only certain "metabolic process" categories
- keep_processes <- c(
- "ATP metabolic process",
- "regulation of lipid metabolic process",
- "superoxide metabolic process"
- )
- df_plot0 <- df_plot0 %>%
- filter(
- # keep all terms that don't end in "metabolic process"
- !stringr::str_detect(Pathway, "metabolic process$") |
- # OR are one of the specific ones you care about
- Pathway %in% keep_processes
- )
- # Keep union of top 10 per target
- n_keep <- 12
- top10_per_treat <- df_plot0 %>%
- group_by(Target) %>%
- arrange(colval, Pathway, .by_group = TRUE) %>% # <- ASC padj
- slice_head(n = n_keep) %>%
- ungroup()
- # Collect the unique GO terms that are top-10 in at least one target
- keepers <- unique(top10_per_treat$Pathway)
- # Filter the *full* table to those GO terms for *all* Treatments where they exist
- df_plot0 <- df_plot0 %>%
- dplyr::filter(Pathway %in% keepers)
- # Y ordering by max -log10(padj) within Group + dividers
- ord <- df_plot0 %>%
- group_by(Group, Pathway) %>%
- summarise(ordkey = suppressWarnings(min(colval, na.rm = TRUE)), .groups = "drop") %>% # <- MIN padj
- mutate(ordkey = ifelse(is.infinite(ordkey), NA_real_, ordkey)) %>%
- group_by(Group) %>%
- arrange(ordkey, Pathway, .by_group = TRUE) %>% # <- ASC padj
- mutate(y_in_group = row_number()) %>%
- ungroup()
- if (nrow(ord) == 0) stop("No rows left to plot after filters and collapsing.")
- sizes <- ord %>%
- dplyr::count(Group, name = "n") %>%
- dplyr::mutate(offset = dplyr::lag(cumsum(n), default = 0))
- ord <- ord %>% left_join(sizes, by = "Group") %>%
- mutate(y = y_in_group + offset)
- y_top <- max(ord$y) + 0.6
- y_bottom <- min(ord$y) - 0.4
- # positions for right-side labels and horizontal dividers
- label_pos <- ord %>%
- dplyr::group_by(Group) %>%
- dplyr::summarise(y = (min(y) + max(y))/2, .groups = "drop")
- # compute one boundary after each group, then drop ONLY the very last boundary by y
- dividers <- ord %>%
- dplyr::group_by(Group) %>%
- dplyr::summarise(y = max(y) + 0.5, .groups = "drop") %>%
- dplyr::filter(y < max(y)) # <- robust: keep all but the largest boundary
- df_plot <- df_plot0 %>%
- left_join(ord %>% dplyr::select(Group, Pathway, y), by = c("Group","Pathway"))
- write.xlsx(df_plot, "analysis/de/go_bp_batch1_padj_0.1_10deg_allselect_rebeccaedit_finalcatstoplot.xlsx")
- # Plot (x order = ΔCLPI-descending; size=count; color=-log10(padj))
- left_tag <- ".LEFT"
- right_tag <- ".RIGHT"
- x_levels <- target_levels
- df_plot <- df_plot %>%
- mutate(Target2 = factor(Target, levels = c(left_tag, x_levels, right_tag)))
- label_pos <- label_pos %>% mutate(x_lab = right_tag)
- p <- ggplot(df_plot, aes(x = Target2, y = y, size = Count, colour = colval)) +
- geom_hline(data = dividers, aes(yintercept = y),
- inherit.aes = FALSE, linewidth = 0.4, colour = "grey70") +
- geom_point(alpha = 0.9) +
- geom_text(data = label_pos, aes(x = x_lab, y = y, label = Group),
- inherit.aes = FALSE, hjust = 0, vjust = 0.5, fontface = "bold") +
- scale_x_discrete(
- drop = FALSE,
- limits = c(left_tag, x_levels, right_tag),
- breaks = x_levels,
- labels = x_levels,
- expand = expansion(add = c(0, 0.65))
- ) +
- scale_y_reverse(
- limits = c(y_top, y_bottom),
- breaks = ord$y,
- labels = ord$Pathway,
- expand = c(0, 0)
- ) +
- scale_size_area(max_size = 11, name = "Gene Count") +
- # clusterProfiler-like red↔blue (red = more significant)
- scale_colour_gradient(
- name = "padj",
- low = "red", # smaller padj = red
- high = "blue",
- oob = scales::squish,
- guide = guide_colorbar(reverse = TRUE) # <- lower numbers appear at the top
- ) +
- labs(x = "Target Gene", y = NULL) +
- coord_cartesian(clip = "off") +
- theme_bw(base_size = 12) +
- theme(
- plot.margin = margin(5.5, 70, 5.5, 5.5),
- axis.text.x = element_text(angle = 45, hjust = 1),
- legend.position = "right"
- )
- p
- # make a dotplot annotated with a boxplot --------
- # Pick a global font family
- base_family <- "Arial"
- # Keep the SAME x limits as dotplot (including sentinel columns)
- x_limits_full <- c(left_tag, x_levels, right_tag)
- # Compute Treatment-level mean (across wells), subtract NEG mean
- .delta_bar_data <- function(data, feature, control = "NEG") {
- # Only keep NEG + treatments that appear in the dotplot order
- present <- intersect(c(control, x_levels), unique(data$Treatment))
- means <- data %>%
- filter(Treatment %in% present) %>%
- group_by(Treatment) %>%
- summarise(mean_val = mean(.data[[feature]], na.rm = TRUE), .groups = "drop")
- neg_mean <- means %>% filter(Treatment == control) %>% pull(mean_val)
- if (length(neg_mean) == 0 || is.na(neg_mean)) {
- stop("NEG control not found in the provided data for feature: ", feature)
- }
- # Delta = gene mean - NEG mean, for non-NEG treatments in the dotplot order
- bars <- tibble(Treatment = x_levels) %>%
- left_join(means, by = "Treatment") %>%
- mutate(delta = mean_val - neg_mean)
- # Build a full x vector including sentinels (no bars for them)
- df <- tibble(Treatment2 = factor(x_limits_full, levels = x_limits_full)) %>%
- left_join(
- bars %>%
- transmute(Treatment2 = factor(Treatment, levels = x_limits_full),
- delta),
- by = "Treatment2"
- ) %>%
- mutate(sign = case_when(
- is.na(delta) ~ "zero", # nothing drawn
- delta > 0 ~ "pos",
- delta < 0 ~ "neg",
- TRUE ~ "zero"
- ))
- df
- }
- # Use Arial in the two delta bar panels
- make_delta_bar_layer <- function(data, feature, panel_title, control = "NEG") {
- df <- .delta_bar_data(data, feature, control = control)
- ggplot(df, aes(x = Treatment2, y = delta, fill = sign)) +
- geom_col(width = 0.7, na.rm = TRUE) +
- geom_hline(yintercept = 0, linewidth = 0.4, color = "grey50") +
- scale_fill_manual(values = c(pos = "red", neg = "blue", zero = "grey80"), guide = "none") +
- scale_x_discrete(drop = FALSE, limits = x_limits_full, breaks = NULL, expand = expansion(add = c(0, 0.65))) +
- labs(title = panel_title, x = NULL, y = NULL) +
- theme_minimal(base_size = 12, base_family = base_family) + # <- Arial
- theme(
- plot.title = element_text(hjust = 0, face = "bold", size = 11, margin = margin(b = 2)),
- axis.text.x = element_blank(),
- axis.title.x = element_blank(),
- axis.ticks.x = element_blank(),
- legend.position = "none",
- plot.margin = margin(5.5, 70, 0, 5.5)
- )
- }
- p <- p +
- theme_bw(base_size = 12, base_family = base_family) + # <- Arial
- theme(plot.margin = margin(5.5, 70, 5.5, 5.5))
- p <- p +
- geom_text(
- data = label_pos,
- aes(x = x_lab, y = y, label = Group),
- inherit.aes = FALSE,
- hjust = 0, vjust = 0.5,
- fontface = "bold",
- family = base_family,
- size = 3
- )
- # Rebuild your two annotation bars with plotmath titles using Δ (delta)
- p_bar_pi <- make_delta_bar_layer(
- exp31, "Cell_Level_Phago_Index",
- panel_title = expression(Delta~" well averaged PI"),
- control = "NEG"
- )
- p_bar_ecc <- make_delta_bar_layer(
- exp31, "AreaShape_Eccentricity",
- panel_title = expression(Delta~" well-averaged Eccentricity"),
- control = "NEG"
- )
- # make sure label_pos has the x column for the right margin labels
- label_pos <- label_pos %>% mutate(x_lab = right_tag)
- # 1) Remove any existing text/label geoms from the dotplot
- strip_text_layers <- function(g) {
- g$layers <- Filter(function(lyr) !inherits(lyr$geom, c("GeomText", "GeomLabel")), g$layers)
- g
- }
- p <- strip_text_layers(p)
- # Add ONE right-side label layer (Arial + smaller size)
- base_family <- "Arial"
- p <- p +
- geom_text(
- data = label_pos,
- aes(x = x_lab, y = y, label = Group),
- inherit.aes = FALSE,
- hjust = 0, vjust = 0.5,
- fontface = "bold",
- family = base_family,
- size = 3.2
- )
- p <- p +
- theme(
- axis.text.x = element_text(
- angle = 45, hjust = 1, vjust = 1,
- family = "Arial", size = 12
- ),
- # give a bit more bottom margin so angled labels don't clip
- plot.margin = margin(5.5, 70, 10, 5.5)
- )
- # combine and save
- combined <- p_bar_pi / p_bar_ecc / p + patchwork::plot_layout(heights = c(0.7, 0.7, 5))
- dir.create("analysis/de", recursive = TRUE, showWarnings = FALSE)
- dev <- if (capabilities("cairo")) grDevices::cairo_pdf else grDevices::pdf
- ggplot2::ggsave(
- filename = "analysis/de/go_final_dotplot_barplot.pdf",
- plot = combined,
- width = 10, height = 14, units = "in",
- device = dev
- )
- # SINGLE CATEGORY gene heatmap ======
- plot_group_heatmaps <- function(df,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Lysosome",
- out_dir = "analysis/plots/single_cat/",
- alpha = 0.1,
- lfc_thresh = 0.2,
- centered_order_tbl = NULL,
- min_genes_after_deg = 1,
- min_treats_after_filter = 1,
- cluster_rows = TRUE,
- cluster_cols = TRUE,
- plot_height = 17,
- plot_width = 5,
- base_family = "Arial",
- lfc_limits = c(-4, 2)
- ) {
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv"
- out_dir = "analysis/plots/single_cat/"
- alpha = 0.1
- lfc_thresh = 0.2
- min_genes_after_deg = 1
- min_treats_after_filter = 1
- cluster_rows = TRUE
- cluster_cols = TRUE
- base_family = "Arial"
- lfc_limits = c(-4, 2)
- df = df_phago
- group = "Phagocytosis"
- centered_order_tbl = centered_order_tbl
- plot_width = sz$width_in
- plot_height = sz$height_in
- # --- Safety & inputs ---
- stopifnot(all(c("select_label","description","gene_id","ko_gene") %in% names(df)))
- if (!dir.exists(out_dir)) dir.create(out_dir, recursive = TRUE, showWarnings = FALSE)
- # Read DEG (only once)
- deg <- data.table::fread(deg_path) %>%
- dplyr::mutate(gene = gene_name)
- # Terms in this group
- terms <- df %>%
- dplyr::filter(select_label == group) %>%
- dplyr::pull(description) %>%
- unique()
- if (length(terms) == 0) stop(sprintf("No terms found with select_label == '%s'.", group))
- # Helper to sanitize filenames
- sanitize <- function(x) gsub("[^A-Za-z0-9_\\-]+", "_", x)
- # PDF device that embeds fonts
- .pdf_device <- if (capabilities("cairo")) grDevices::cairo_pdf else grDevices::pdf
- # Plot each term
- out_files <- list()
- for (term_desc in terms) {
- message("Plotting [", group, "]: ", term_desc)
- # genes for this term
- genes <- df %>%
- dplyr::filter(select_label == group, description == term_desc) %>%
- dplyr::pull(gene_id) %>%
- stringr::str_split("/", simplify = FALSE) %>%
- unlist(use.names = FALSE) %>%
- stringr::str_trim() %>%
- unique()
- genes <- genes[genes != ""]
- if (!length(genes)) { warning("No genes parsed for: ", term_desc); next }
- # treatments that actually list this term in df
- treats_in_term_df <- df %>%
- dplyr::filter(select_label == group, description == term_desc) %>%
- dplyr::pull(ko_gene) %>%
- unique()
- if (!length(treats_in_term_df)) { warning("No treatments listed for this term in df: ", term_desc); next }
- # subset DEG
- deg_sel <- deg %>%
- dplyr::filter(gene %in% genes) %>%
- dplyr::select(gene, log2FoldChange, padj, ko_gene)
- if (!nrow(deg_sel)) { warning("No DEG rows matched for: ", term_desc); next }
- # wide -> matrices
- lfc_wide <- deg_sel %>%
- dplyr::select(gene, ko_gene, log2FoldChange) %>%
- tidyr::pivot_wider(names_from = ko_gene, values_from = log2FoldChange)
- p_wide <- deg_sel %>%
- dplyr::select(gene, ko_gene, padj) %>%
- tidyr::pivot_wider(names_from = ko_gene, values_from = padj)
- lfc_raw <- lfc_wide %>% tibble::column_to_rownames("gene") %>% as.matrix()
- p_raw <- p_wide %>% tibble::column_to_rownames("gene") %>% as.matrix()
- # drop cols not present in df for this term
- lfc_raw <- lfc_raw[, colnames(lfc_raw) %in% treats_in_term_df, drop = FALSE]
- p_raw <- p_raw[, colnames(p_raw) %in% treats_in_term_df, drop = FALSE]
- # drop all-NA columns, then order by ΔCLPI (only remaining)
- nonempty_cols <- colSums(!is.na(lfc_raw)) > 0
- lfc_raw <- lfc_raw[, nonempty_cols, drop = FALSE]
- p_raw <- p_raw[, nonempty_cols, drop = FALSE]
- desired_order <- intersect(centered_order_tbl$Treatment, colnames(lfc_raw))
- lfc_raw <- lfc_raw[, desired_order, drop = FALSE]
- p_raw <- p_raw[, desired_order, drop = FALSE]
- if (ncol(lfc_raw) < min_treats_after_filter) { warning("Too few treatments after filtering for: ", term_desc); next }
- # fill NA (after pruning)
- lfc_mat <- lfc_raw; lfc_mat[is.na(lfc_mat)] <- 0
- p_mat <- p_raw; p_mat[is.na(p_mat)] <- 1
- # drop genes with no data across remaining treatments
- nonempty_rows <- rowSums(!is.na(lfc_raw)) > 0
- lfc_mat <- lfc_mat[nonempty_rows, , drop = FALSE]
- p_mat <- p_mat[nonempty_rows, , drop = FALSE]
- if (nrow(lfc_mat) < min_genes_after_deg) { warning("Too few genes with data left for: ", term_desc); next }
- heatmap_max <- suppressWarnings(max(lfc_mat, na.rm = TRUE)); if (!is.finite(heatmap_max)) heatmap_max <- 1
- heatmap_min <- suppressWarnings(min(lfc_mat, na.rm = TRUE)); if (!is.finite(heatmap_min)) heatmap_min <- -1
- # star overlay (use Arial)
- cell_fun_star <- function(j, i, x, y, w, h, fill) {
- if (!is.na(p_mat[i, j]) && p_mat[i, j] < alpha &&
- !is.na(lfc_mat[i, j]) && abs(lfc_mat[i, j]) > lfc_thresh) {
- grid::grid.text("*", x, y, gp = grid::gpar(fontfamily = base_family))
- }
- }
- heatmap_min <- lfc_limits[1]
- heatmap_max <- lfc_limits[2]
- # keep white at 0 (diverging)
- col_fun <- circlize::colorRamp2(
- c(heatmap_min, 0, heatmap_max),
- c("darkblue", "white", "red")
- )
- # legend ticks (top→bottom shows 2, 0, -2, -4). If yours renders bottom→top,
- # swap to at = c(-4, -2, 0, 2) to match your preference.
- legend_at <- c(2, 0, -2, -4)
- legend_labels <- c("2", "0", "-2", "-4")
- pdf_file <- file.path(out_dir, paste0(sanitize(group), "__", sanitize(term_desc), ".pdf"))
- .pdf_device(pdf_file, height = plot_height, width = plot_width, family = base_family)
- view(lfc_mat)
- ht <- ComplexHeatmap::Heatmap(
- lfc_mat,
- name = "Log2FC",
- col = col_fun,
- cluster_columns = cluster_cols,
- cluster_rows = cluster_rows,
- column_names_gp = grid::gpar(fontsize = 10, fontfamily = base_family),
- row_names_gp = grid::gpar(fontsize = 10, fontfamily = base_family),
- heatmap_legend_param = list(
- title = "Log2FoldChange\nKO vs. Control",
- title_gp = grid::gpar(fontfamily = base_family, fontface = "bold"),
- labels_gp = grid::gpar(fontfamily = base_family),
- at = legend_at,
- labels = legend_labels
- ),
- column_title = term_desc,
- column_title_gp = grid::gpar(fontfamily = base_family, fontface = "bold"),
- cell_fun = cell_fun_star
- )
- ComplexHeatmap::draw(ht)
- grDevices::dev.off()
- message("Saved: ", pdf_file)
- out_files[[term_desc]] <- pdf_file
- }
- invisible(out_files)
- }
- estimate_heatmap_dims <- function(df_plot, deg_path, group, centered_order_tbl) {
- # genes for this group/term set (split "gene_id" on '/')
- genes <- df_plot %>%
- filter(select_label == group) %>%
- pull(gene_id) %>%
- str_split("/", simplify = FALSE) %>%
- unlist(use.names = FALSE) %>%
- str_trim() %>%
- unique()
- genes <- genes[genes != ""]
- # treatments mentioned for these terms
- treats_in_term_df <- df_plot %>%
- filter(select_label == group) %>%
- pull(ko_gene) %>%
- unique()
- # read DEG and subset to these genes/treatments
- deg <- data.table::fread(deg_path) %>% mutate(gene = gene_name)
- deg_sel <- deg %>%
- filter(gene %in% genes, ko_gene %in% treats_in_term_df) %>%
- select(gene, ko_gene, log2FoldChange)
- # wide, then drop all-NA cols and keep centered order
- lfc_wide <- deg_sel %>%
- pivot_wider(names_from = ko_gene, values_from = log2FoldChange)
- if (nrow(lfc_wide) == 0) return(list(nr = 0L, nc = 0L))
- lfc_raw <- lfc_wide %>% column_to_rownames("gene") %>% as.matrix()
- nonempty_cols <- colSums(!is.na(lfc_raw)) > 0
- lfc_raw <- lfc_raw[, nonempty_cols, drop = FALSE]
- desired_order <- intersect(centered_order_tbl$Treatment, colnames(lfc_raw))
- lfc_raw <- lfc_raw[, desired_order, drop = FALSE]
- # drop genes with all NA across remaining treatments
- nonempty_rows <- rowSums(!is.na(lfc_raw)) > 0
- lfc_raw <- lfc_raw[nonempty_rows, , drop = FALSE]
- list(nr = nrow(lfc_raw), nc = ncol(lfc_raw))
- }
- heatmap_device_size <- function(nr, nc,
- cell_w_mm = 6,
- cell_h_mm = 6.5,
- row_names_mm = 35,
- col_names_mm = 12,
- row_dend_mm = 4,
- col_dend_mm = 4,
- legend_mm = 18,
- gap_mm = 6) {
- mm_to_in <- function(mm) mm / 25.4
- heatmap_w_in <- mm_to_in(cell_w_mm * nc)
- heatmap_h_in <- mm_to_in(cell_h_mm * nr)
- width_in <- heatmap_w_in + mm_to_in(row_names_mm + row_dend_mm + gap_mm + legend_mm)
- height_in <- heatmap_h_in + mm_to_in(col_names_mm + col_dend_mm + gap_mm)
- list(width_in = width_in, height_in = height_in)
- }
- # Example: Phagocytosis panel
- df_phago <- df %>% filter(description == "phagocytosis")
- head(df_phago)
- dims <- estimate_heatmap_dims(df_phago, deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Phagocytosis", centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_phago <- plot_group_heatmaps(
- df_phago,
- group = "Phagocytosis",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- # Lysosome terms
- df_plot = df %>%
- filter(description == "lysosome organization")
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Lysosome",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_lyso <- plot_group_heatmaps(
- df_plot,
- group = "Lysosome",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- # Activation terms
- df_plot = df %>%
- filter(description == "macrophage activation")
- head(df_plot)
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Inflammation/Activation",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_act <- plot_group_heatmaps(
- df_plot,
- group = "Inflammation/Activation",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- # Actin terms
- df_plot = df %>%
- filter(description == "regulation of cell shape")
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_act <- plot_group_heatmaps(
- df_plot,
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- # Actin term
- df_plot = df %>%
- filter(description == "actin crosslink formation")
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_act <- plot_group_heatmaps(
- df_plot,
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- df_plot = df %>%
- filter(description == "actin filament bundle assembly")
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_act <- plot_group_heatmaps(
- df_plot,
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
- df_plot = df %>%
- filter(description == "regulation of actin cytoskeleton organization")
- dims <- estimate_heatmap_dims(df_plot,
- deg_path = "analysis/de/all_res_vs_ntc1_batch1_q1q3_filter_rmlow.tsv",
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl)
- sz <- heatmap_device_size(dims$nr, dims$nc)
- files_act <- plot_group_heatmaps(
- df_plot,
- group = "Actin/Cell Shape",
- centered_order_tbl = centered_order_tbl,
- plot_width = sz$width_in,
- plot_height = sz$height_in
- )
03_01_analyze_deg_go.R at commit 7ea1019, no license · at the source
Overview
- Center for Genomic Medicine and Department of Psychiatry, Massachusetts General Hospital,Boston, MA USA
- Department of Molecular Biology, Massachusetts General Hospital,Boston, MA USA
- Tufts Institute for Artificial Intelligence, Tufts University,Medford, MA USA
- Department of Psychiatry, Harvard Medical School,Boston, MA USA
- Harvard Stem Cell Institute,Cambridge, MA USA
Abstract
Microglia are increasingly recognized as key regulators of neural circuit development and putative contributors to the pathophysiology of neuropsychiatric disorders such as schizophrenia (SCZ). However, the functional impact of SCZ-associated genes in microglia remains largely unexplored. Here, we performed an arrayed CRISPR targeting screen of 30 SCZ-associated genes predicted to be differentially expressed in human microglia-like cells. Target genes were prioritized based on post-mortem transcriptomic relevance and predicted ontology-based roles in phagocytosis pathways. We quantified phagocytic activity and morphological changes following gene targeting using high-content confocal imaging. Key targets, including CYFIP1, MSR1, TREM2, SYK, ITGB2, ITGAM, and IRF8, modulated phagocytosis and altered morphological properties consistent with activation states, validating their functional roles in microglia. To elucidate transcriptional impact, we further applied a multiplexed RNA sequencing platform across gene targets. These analyses revealed gene-specific transcriptional signatures, implicating divergent pathways related to phagocytic, activation, cytoskeletal, and lysosomal function. Together, these findings demonstrate the utility of CRISPR-based functional genomics in characterizing microglia function and identifying new target genes and mechanisms that may underlie their contributions to SCZ pathophysiology.
Reproduced under the paper's license (CC BY), from the paper cited above.
Repository
Its files are read in the Code ↔ Paper reader above, with 7 matches between paragraphs and lines of code.
mgb-cqh/crispr30
7ea101924cd85986ebddc043eb9d661e32c241b3, 21 November 2025Availability: 1 check, the latest on 29 September 2026: the link answers
- 29 September 2026: the link answers
12 files
- drug_seq/
00_00_sbatch_trim_fastqc , Shell, 19 lines.sh - drug_seq/
01_00_sbatch_align.sh , Shell, 44 lines - drug_seq/
01_01_sbatch_bam_demulti , Shell, 51 linesplex.sh - drug_seq/
01_02_run_sbatch_bam_dem , Shell, 18 linesultiplex.sh - drug_seq/
01_03_sbatch_compute_sat , Shell, 47 linesuration.sh - drug_seq/
02_00_generate_count_mat , R, 112 linesrix.R - drug_seq/
02_01_analyze_count_matr , R, 213 linesix.R - drug_seq/
03_00_deg.R , R, 191 lines - drug_seq/
03_01_analyze_deg_go.R , R, 1,454 lines, 3 matches - drug_seq/
generate_string_plot_fin , R, 305 lines, 2 matchesal.Rmd - image_analysis/
Crispr30_publication_fin , R, 1,094 lines, 2 matchesal.Rmd - README.md, Text, 2 lines
Tracing map
Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.
What the map holds:
- 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
- 11 scripts, each with its path and the digest of its content;
- 7 matches between paragraphs of the paper and lines of the code (method lexical-v1);
- neither the text of the paper nor the code itself.
Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.
Data
Datasets cited
- geo:GSE319209, at NCBI GEO; found in “Data availability”
Data availability
Gene expression data will be made available for download from the NCBI Gene Expression Omnibus (https://
Reproduced under the paper's license (CC BY), from the paper cited above.
Versions
The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.
Version 1, 29 September 2026: the first record
Recorded: type, language, journal, volume, issue, pages, dates, 8 authors, 2 keywords, 6 MeSH terms, 1 funder, 74 references.
Cite
This paper
Horng, J. E., McCrea, L. T., Batorsky, R. E., Bowen, J. J., Boschian, C., Song, Y., Perlis, R. H., & Sheridan, S. D. (2026). Functional genomic profiling of schizophrenia-associated
BibTeX
@article{horng2026functi
author = {Horng, Joy E. and McCrea, Liam T. and Batorsky, Rebecca E. and Bowen, Joshua J. and Boschian, Camilla and Song, Yoonjae and Perlis, Roy H. and Sheridan, Steven D.},
title = {{Functional genomic profiling of schizophrenia-associated
journal = {Neuropsychopharmacology
year = {2026},
month = apr,
volume = {51},
number = {9},
pages = {1718--1727},
publisher = {Nature Publishing Group},
issn = {0893-133X},
doi = {10.1038/
url = {https://
pmid = {41992032},
pmcid = {PMC13389400}
}
RIS
TY - JOUR
AU - Horng, Joy E.
AU - McCrea, Liam T.
AU - Batorsky, Rebecca E.
AU - Bowen, Joshua J.
AU - Boschian, Camilla
AU - Song, Yoonjae
AU - Perlis, Roy H.
AU - Sheridan, Steven D.
TI - Functional genomic profiling of schizophrenia-associated
T2 - Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology
J2 - Neuropsychopharmacology
PY - 2026
DA - 2026/
VL - 51
IS - 9
SP - 1718
EP - 1727
SN - 0893-133X
PB - Nature Publishing Group
DO - 10.1038/
UR - https://
LA - en
ER -
CSL-JSON
{
"id": "10.1038/
"type": "article-journal",
"title": "Functional genomic profiling of schizophrenia-associated
"container-title": "Neuropsychopharmacology
"author": [
{
"family": "Horng",
"given": "Joy E."
},
{
"family": "McCrea",
"given": "Liam T."
},
{
"family": "Batorsky",
"given": "Rebecca E."
},
{
"family": "Bowen",
"given": "Joshua J."
},
{
"family": "Boschian",
"given": "Camilla"
},
{
"family": "Song",
"given": "Yoonjae"
},
{
"family": "Perlis",
"given": "Roy H."
},
{
"family": "Sheridan",
"given": "Steven D."
}
],
"container-title-short":
"volume": "51",
"issue": "9",
"page": "1718-1727",
"DOI": "10.1038/
"PMID": "41992032",
"PMCID": "PMC13389400",
"ISSN": "0893-133X",
"publisher": "Nature Publishing Group",
"URL": "https://
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
16
]
]
}
}
The tracing map gets a citation of its own once an author has validated it and it has a DOI.
Similar papers
The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.
- [1] doi:10.1038/s41386-025-02123-1
- Identifying brain-penetrant small-molecule modulators of human microglia using a cellular model of synaptic pruning.Journal: Neuropsychopharmacology : official publication of the American College of NeuropsychopharmacologyIn common: cellular / molecular, 17 references, 3 authors
- [2] doi:10.1186/s12974-024-03157-w [code]
- Maternal SARS-CoV-2 impacts fetal placental macrophage programs and placenta-derived microglial models of neurodevelopment.Journal: Journal of neuroinflammationIn common: circlize, clusterProfiler, ComplexHeatmap, 7 other tools, genetics / omics, cellular / molecular, 7 references, 2 authors
- [3] doi:10.1093/bioinformatics/btag592 [code]
- Network-based stratification of allele-specific expression reveals patient subgroups in Huntington's disease.Journal: Bioinformatics (Oxford, England)In common: FastQC, multcomp, STAR, 16 other tools, genetics / omics, 4 references
- [4] doi:10.1186/s12974-024-03198-1 [code]
- Repeated LPS induces training and tolerance of microglial responses across brain regions.Journal: Journal of neuroinflammationIn common: FastQC, STAR, SAMtools, 10 other tools, cellular / molecular, 6 references
- [5] doi:10.1038/s41588-021-00976-y [code]
- Genetic analysis of the human microglial transcriptome across brain regions, aging and disease pathologies.Journal: Nature geneticsIn common: SAMtools, broom, limma, 10 other tools, genetics / omics, cellular / molecular, 7 references
- [6] doi:10.1016/j.celrep.2024.114326 [code]
- Hofbauer cells and fetal brain microglia share transcriptional profiles and responses to maternal diet-induced obesity.Journal: Cell reportsIn common: limma, circlize, clusterProfiler, 7 other tools, genetics / omics, cellular / molecular, 4 references, author Roy H. Perlis
- [7] doi:10.1038/s43587-025-00816-2 [code]
- Interleukin-12 signaling drives Alzheimer's disease pathology through disrupting neuronal and oligodendrocyte homeostasis.Journal: Nature agingIn common: multcomp, STAR, SAMtools, 12 other tools, cellular / molecular, 2 references
- [8] doi:10.1038/s41593-026-02367-0 [code]
- A reproducible three-dimensional model of human brain tissue to investigate physiological and disease-associated microglia phenotypes.Journal: Nature neuroscienceIn common: multcomp, limma, circlize, 10 other tools, cellular / molecular, 4 references
- [9] doi:10.1038/s41467-023-41033-y [code]
- Divergent single cell transcriptome and epigenome alterations in ALS and FTD patients with C9orf72 mutation.Journal: Nature communicationsIn common: FastQC, STAR, SAMtools, 11 other tools, genetics / omics, cellular / molecular, 2 references
- [10] doi:10.1126/sciadv.ade2812 [code]
- Consensus molecular environment of schizophrenia risk genes in coexpression networks shifting across age and brain regions.Journal: Science advancesIn common: broom, limma, circlize, 9 other tools, schizophrenia / psychosis, 5 references
Contribute
The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.
Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.
Claim this paper
Correct its record
Say what each link of this record is, remove the ones that are not the paper's, add the ones that are missing. The correction becomes a new version of the record, in its Versions section.
Validate its tracing map
You validate the map as this page shows it: 1 repository of the authors' code, each at its verified commit and with its license, 11 scripts, and 7 matches between paragraphs and code (see the Code and Map sections). It then receives a DOI on Zenodo, with you (your ORCID iD) and OSCR as its creators; the code itself is not deposited.
The map's fingerprint: sha256:7da6fffe0a9a7892…
Add the badge to the code's README
If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.
Markdown
<!-- oscr-badge:start -->
[](https://openscicode.org/badge/doi/10.1038/s41386-026-02406-1)
<!-- oscr-badge:end -->reStructuredText
.. oscr-badge:start
.. image:: https://openscicode.org/badge/doi/10.1038/s41386-026-02406-1.svg
:target: https://openscicode.org/badge/doi/10.1038/s41386-026-02406-1
:alt: OSCR
.. oscr-badge:endHTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.1038/s41386-026-02406-1"><img src="https://openscicode.org/badge/doi/10.1038/s41386-026-02406-1.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".
If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge page.
Request its removal
To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).
Discussion, reproductions, activity
Discussion: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.
Report a mismatch between the code and the paper · Report a code error (signed in)
Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.
Report a reproduction failure (signed in)
Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.
Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.
