8. Protein — Intensity profiles

Overview

Per-protein intensity profiles for all significant DE proteins (FDR < 0.05, |log2FC| > 0.5) in each contrast. Each panel shows the MSstats-summarised log₂ intensity across the four conditions, with individual replicate points and a mean ± SE crossbar. Panels are ordered by adjusted p-value (most significant first) and paginated at 20 proteins per page.

Inputs:

• Full_GroupComparison_Results.csv — MSstats DE results from notebook 02
• processed_msstats_data.qs2 → $ProteinLevelData — per-run summarised log₂ intensities
• protein_dictionary.csv — UniProt → gene symbol mapping built in notebook 03

Thresholds match notebook 03 (volcano): FDR < 0.05, |log2FC| > 0.5.

Libraries

Directories

base_dir <- "/nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics"
run_num  <- "run4"

results_dir <- file.path(base_dir, "results", run_num)
objects_dir <- file.path(base_dir, "data", "processed", run_num)

Condition palette and order

Same 2×2 factorial colour scheme as the phospho profile plots (notebook 06b).

COND_ORDER <- c("hApp", "hAppNLGF", "hApp_FIRE", "hAppNLGF_FIRE")

cond_cols <- c(
  hApp          = "#2166AC",
  hAppNLGF      = "#D6604D",
  hApp_FIRE     = "#74ADD1",
  hAppNLGF_FIRE = "#F4A582"
)

cond_labels <- c(
  hApp          = "hApp",
  hAppNLGF      = "hApp\nNLGF",
  hApp_FIRE     = "hApp\nFIRE",
  hAppNLGF_FIRE = "hApp\nNLGF\nFIRE"
)

Load DE results and gene symbol mapping

de <- fread(file.path(results_dir, "Full_GroupComparison_Results.csv"))

prot_dict <- fread(file.path(objects_dir, "protein_dictionary.csv"),
                   select = c("Protein", "Gene"))
prot_dict <- unique(prot_dict)

de <- merge(de, prot_dict, by = "Protein", all.x = TRUE)
de[, Gene := ifelse(is.na(Gene) | Gene == "", Protein, Gene)]

ADJ_PVAL_CUTOFF <- 0.05
LOGFC_CUTOFF    <- 0.5

de_sig <- de[adj.pvalue < ADJ_PVAL_CUTOFF & abs(log2FC) > LOGFC_CUTOFF]
cat("Significant proteins per contrast:\n")

Significant proteins per contrast:

print(de_sig[, .N, by = Label][order(Label)])

                         Label     N
                        <char> <int>
1: Interaction_APP_x_Microglia    40
2:   hAppNLGF_FIRE_vs_hAppNLGF    81
3:  hAppNLGF_FIRE_vs_hApp_FIRE    33
4:            hAppNLGF_vs_hApp    31
5:           hApp_FIRE_vs_hApp    99

Load summarised intensities

summ <- qs_read(file.path(objects_dir, "processed_msstats_data.qs2"))
pld  <- as.data.table(summ$ProteinLevelData)
pld[, GROUP := factor(GROUP, levels = COND_ORDER)]
cat("ProteinLevelData rows:", nrow(pld),
    " | unique proteins:", uniqueN(pld$Protein), "\n")

ProteinLevelData rows: 140794  | unique proteins: 8890 

Profile-plot helper

PAGE_SIZE <- 20

make_profiles <- function(contrast_label) {
  sig_ct <- de_sig[Label == contrast_label][order(adj.pvalue)]
  n_sig  <- nrow(sig_ct)
  if (n_sig == 0) {
    message("No significant proteins for ", contrast_label, " — skipping.")
    return(invisible(NULL))
  }

  sig_ct[, FacetLabel := sprintf(
    "%s\npadj=%.3g | FC=%.2f", Gene, adj.pvalue, log2FC
  )]
  sig_ct[, FacetLabel := factor(FacetLabel, levels = unique(FacetLabel))]

  pld_sig <- pld[Protein %in% sig_ct$Protein]
  pld_sig <- merge(pld_sig,
                   sig_ct[, .(Protein, FacetLabel)],
                   by = "Protein")
  pld_sig[, FacetLabel := factor(FacetLabel, levels = levels(sig_ct$FacetLabel))]

  n_pages  <- ceiling(n_sig / PAGE_SIZE)
  safe     <- gsub(" ", "_", contrast_label)
  pdf_path <- file.path(results_dir, glue("ProteinProfiles_{safe}.pdf"))

  plots <- vector("list", n_pages)

  cairo_pdf(pdf_path, onefile = TRUE, width = 14, height = 11)
  for (pg in seq_len(n_pages)) {
    idx_lo   <- (pg - 1L) * PAGE_SIZE + 1L
    idx_hi   <- min(pg * PAGE_SIZE, n_sig)
    pg_sites <- levels(sig_ct$FacetLabel)[idx_lo:idx_hi]
    pg_data  <- pld_sig[FacetLabel %in% pg_sites]
    pg_data[, FacetLabel := factor(FacetLabel, levels = pg_sites)]

    ncol_n       <- min(5L, length(pg_sites))
    page_caption <- if (n_pages > 1) glue("Page {pg} of {n_pages}") else NULL

    p <- ggplot(pg_data,
                aes(x = GROUP, y = LogIntensities, colour = GROUP)) +
      stat_summary(fun.data = mean_se,
                   geom      = "crossbar",
                   mapping   = aes(fill = GROUP),
                   width     = 0.55,
                   linewidth = 0.45,
                   fatten    = 1,
                   alpha     = 0.25,
                   show.legend = FALSE) +
      geom_jitter(width = 0.12, size = 1.8, alpha = 0.9) +
      scale_colour_manual(values = cond_cols, limits = COND_ORDER) +
      scale_fill_manual(values   = cond_cols, limits = COND_ORDER) +
      scale_x_discrete(limits = COND_ORDER, labels = cond_labels) +
      facet_wrap(~ FacetLabel, ncol = ncol_n, scales = "free_y") +
      labs(
        x        = NULL,
        y        = "Summarised log₂ intensity",
        title    = glue("Significant proteins — {gsub('_', ' ', contrast_label)}"),
        subtitle = glue(
          "adj.p < {ADJ_PVAL_CUTOFF}  |  |log₂FC| > {LOGFC_CUTOFF}  ",
          "({n_sig} protein{ifelse(n_sig==1,'','s')})"
        ),
        caption = page_caption,
        colour  = "Condition"
      ) +
      theme_minimal(base_size = 10) +
      theme(
        plot.title           = element_text(face = "bold", size = 13, hjust = 0),
        plot.subtitle        = element_text(size = 9, colour = "grey40",
                                            margin = margin(b = 8)),
        strip.text           = element_text(face = "bold", size = 7.5,
                                            lineheight = 1.15),
        axis.text.x          = element_text(size = 7, lineheight = 0.9),
        axis.text.y          = element_text(size = 7),
        axis.title.y         = element_text(size = 9),
        panel.grid.minor     = element_blank(),
        panel.grid.major.x   = element_blank(),
        panel.border         = element_rect(colour = "grey85", fill = NA,
                                            linewidth = 0.4),
        legend.position      = "top",
        legend.title         = element_text(size = 9, face = "bold"),
        legend.text          = element_text(size = 8)
      ) +
      guides(colour = guide_legend(nrow = 1, override.aes = list(size = 3)))

    plots[[pg]] <- p
    print(p)  # → PDF device
  }
  dev.off()  # close PDF before printing to screen

  for (p in plots) print(p)

  cat("Saved:", pdf_path, "(", n_pages, "page(s),", n_sig, "proteins)\n")
  invisible(pdf_path)
}

Generate profile plots — all contrasts

contrasts_all <- sort(unique(de_sig$Label))
for (ct in contrasts_all) make_profiles(ct)

Saved: /nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics/results/run4/ProteinProfiles_hApp_FIRE_vs_hApp.pdf ( 5 page(s), 99 proteins)

Saved: /nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics/results/run4/ProteinProfiles_hAppNLGF_FIRE_vs_hApp_FIRE.pdf ( 2 page(s), 33 proteins)

Saved: /nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics/results/run4/ProteinProfiles_hAppNLGF_FIRE_vs_hAppNLGF.pdf ( 5 page(s), 81 proteins)

Saved: /nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics/results/run4/ProteinProfiles_hAppNLGF_vs_hApp.pdf ( 2 page(s), 31 proteins)

Saved: /nemo/lab/destrooperb/home/shared/zanettc/giulia_proteomics/bulk_proteomics/results/run4/ProteinProfiles_Interaction_APP_x_Microglia.pdf ( 2 page(s), 40 proteins)