This document includes the code used for the manuscript, for the transcriptional signature identification.

counts_scaled =
  counts_cell_type %>%

    # Convert to tidybulk tibble
  tidybulk(sample, symbol, count) %>%

    # Preprocess and scale the data
  aggregate_duplicates() %>%
  identify_abundant() %>%
  scale_abundance() %>%

    # Impute missing sample-transcript pairs
    impute_missing_abundance(~cell_type) %>%
    mutate(.abundant = TRUE)
counts_non_red =
  counts_scaled %>%

    # Perform operation for each cell type
    nest(data = -cell_type) %>%
    mutate(data = map(
        data,
        ~ .x %>%
            remove_redundancy(
            method="correlation",
            correlation_threshold = 0.99,
            top=1000
         )
    )) %>%
    unnest(data)
# Select genes that are in at least one sample for all cell types
gene_all = 
    counts_non_red %>%
        distinct(symbol, cell_type) %>%
    count(symbol) %>%
    filter(n == max(n))

# filter dataset and impute missing transcripts-samples pairs
counts_non_red_common =
    counts_non_red %>%
    inner_join(gene_all) 
counts_non_red_common %>%
  reduce_dimensions(method = "tSNE", action="get") %>%
  ggplot(aes(x = `tSNE1`, y = `tSNE2`, color = cell_type)) +
  geom_point(size =2)

plot of chunk unnamed-chunk-7

markers =

  # Define all-versus-all cell type permutations
  counts_non_red_common %>%
  distinct(cell_type) %>%
  pull(cell_type) %>%
  gtools::permutations(n = length(.), r = 2, v = .) %>%
  as_tibble() %>%
  setNames(c("cell_type1", "cell_type2")) %>%
  mutate(contrast = sprintf("cell_type%s - cell_type%s", cell_type1, cell_type2)) %>%

  # Rank marker genes
  mutate(de =
    pmap(
      list(cell_type1, cell_type2, contrast),
      ~   counts_non_red_common %>%
        filter(cell_type %in% c(..1, ..2)) %>%
        test_differential_abundance(~ 0 + cell_type, .contrasts = ..3, fill_missing_values = TRUE, action="get", omit_contrast_in_colnames = T) %>%
        filter(logFC > 0) %>%
        arrange(FDR) %>%
       rowid_to_column(var = "i")
    )) %>%
  unnest(de)
markers %>%

    # Filter best markers for monocytes
    filter(cell_type1=="monocyte" & i==1) %>%

    # Prettify contrasts for plotting
    unite(pair, c("cell_type1", "cell_type2"), remove = FALSE, sep = "\n") %>%

    # Reshape
    gather(which, cell_type, cell_type1, cell_type2) %>%
    distinct(pair,  symbol,   which, cell_type) %>%

    # Attach counts
    left_join(counts_non_red) %>%

    # Plot
    ggplot(aes(y = count_scaled + 1, x = cell_type, fill = cell_type)) +
    geom_boxplot() +
    facet_wrap(~pair+ symbol, scales ="free_x", nrow = 2) +
    scale_y_log10()

plot of chunk unnamed-chunk-10

markers %>%

  # Select first 5 markers from each cell-type pair
  filter(i <= 5) %>%
  unite(pair, c("cell_type1", "cell_type2"), remove = FALSE, sep = "\n") %>%

  # Reshape
  gather(which, cell_type, cell_type1, cell_type2) %>%
  distinct(symbol) %>%

  # Attach counts
  left_join(counts_non_red, by = c("symbol"))  %>%

  # Plot
  reduce_dimensions(sample, symbol, count_scaled, method = "tSNE", action="get") %>%
  pivot_sample(sample) %>%
  ggplot(aes(x = `tSNE1`, y = `tSNE2`, color = cell_type)) +
  geom_point(size =2) 

plot of chunk unnamed-chunk-12