+
+
+
+
+
+
+
+
+
+
+Summary of cell type ontologies in +reference files
+Ally Hawkins
+2024-12-18
+ +
+
+
+
+This notebook aims to identify a set of consensus labels between cell +types in the PanglaoDB and Blueprint Encode references.
+-
+
- First I look at the entire cell type ontology and summarize the +total ancestors and descendants. +
- Then we find all possible combinations of cell type labels between +PanglaoDB and Blueprint Encode and find the latest common ancestor (LCA) +for all combinations. +
- The total descendants for each LCA is used to define a cutoff for +consensus terms we may want to use. +
- I also explored the pairs for some terms in depth to look at +borderline LCA terms. +
- Finally, I calculated the similarity index for all pairs. +
+
+Setup
+suppressPackageStartupMessages({
+ # load required packages
+ library(ggplot2)
+})
+
+# Set default ggplot theme
+theme_set(
+ theme_bw()
+)# The base path for the OpenScPCA repository, found by its (hidden) .git directory
+repository_base <- rprojroot::find_root(rprojroot::is_git_root)
+
+# The path to this module
+ref_dir <- file.path(repository_base, "analyses", "cell-type-consensus", "references")
+
+# path to ref file for panglao
+panglao_file <- file.path(ref_dir, "panglao-cell-type-ontologies.tsv")# grab obo file
+cl_ont <- ontologyIndex::get_ontology("http://purl.obolibrary.org/obo/cl-basic.obo")
+
+# read in panglao file
+panglao_df <- readr::read_tsv(panglao_file) |>
+ # rename columns to have panglao in them for easy joining later
+ dplyr::select(
+ panglao_ontology = "ontology_id",
+ panglao_annotation = "human_readable_value"
+ )## Rows: 178 Columns: 3
+## ── Column specification ──────────────────────────────────────────────────────────────────────────────────────────────────────
+## Delimiter: "\t"
+## chr (3): ontology_id, human_readable_value, panglao_cell_type
+##
+## ℹ Use `spec()` to retrieve the full column specification for this data.
+## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.# grab singler ref from celldex
+blueprint_ref <- celldex::BlueprintEncodeData()
+# get ontologies and human readable name into data frame
+blueprint_df <- data.frame(
+ blueprint_ontology = blueprint_ref$label.ont,
+ blueprint_annotation_main = blueprint_ref$label.main,
+ blueprint_annotation_fine = blueprint_ref$label.fine
+) |>
+ unique()
+
+Full cell ontology
+Below I will calculate the total number of ancestors and the total +number of descendants for each term in the full cell type ontology and +then show the distributions for those statistics. This will give us an +idea of the range of values we expect to see when looking at the +PanglaoDB and Blueprint Encode references.
+# turn cl_ont into data frame with one row per term
+cl_df <- data.frame(
+ cl_ontology = cl_ont$id,
+ cl_annotation = cl_ont$name
+) |>
+ dplyr::rowwise() |>
+ dplyr::mutate(
+ # list all ancestors and descendants calculate total
+ ancestors = list(ontologyIndex::get_ancestors(cl_ont, cl_ontology)),
+ total_ancestors = length(ancestors),
+ descendants = list(ontologyIndex::get_descendants(cl_ont, cl_ontology, exclude_roots = TRUE)),
+ total_descendants = length(descendants)
+ )The vertical lines in the below plot indicate the value for cell +types of varying granularity.
+celltypes_of_interest <- c("eukaryotic cell", "lymphocyte", "leukocyte", "hematopoietic cell", "T cell", "endothelial cell", "smooth muscle cell", "memory T cell")
+line_df <- cl_df |>
+ dplyr::filter(cl_annotation %in% celltypes_of_interest) |>
+ dplyr::select(cl_annotation, total_descendants, total_ancestors) |>
+ unique()
+
+# group any labels that have the same number of ancestors
+ancestor_labels_df <- line_df |>
+ dplyr::group_by(total_ancestors) |>
+ dplyr::summarise(cl_annotation = paste(cl_annotation, collapse = ","))# make density plots showing distribution of ancestors and descendants
+ggplot(cl_df, aes(x = total_ancestors)) +
+ geom_density(fill = "#00274C", alpha = 0.5) +
+ geom_vline(data = ancestor_labels_df,
+ mapping = aes(xintercept = total_ancestors),
+ lty = 2) +
+ geom_text(
+ data = ancestor_labels_df,
+ mapping = aes(x = total_ancestors, y = 0.04, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ labs(
+ x = "Number of ancestors",
+ y = "Density"
+ )
Generally it looks like as the cell types get more specific we see a +greater number of ancestors. However, the range of values is small and +we see some cell types have the same value and probably not the same +level of granularity.
+Below we will look at total number of descendants.
+ggplot(cl_df, aes(x = total_descendants)) +
+ geom_density(fill = "#FFCB05", alpha = 0.5) +
+ geom_vline(data = line_df,
+ mapping = aes(xintercept = total_descendants),
+ lty = 2) +
+ geom_text(
+ data = line_df,
+ mapping = aes(x = total_descendants, y = 0.6, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ labs(
+ x = "Number of descendants",
+ y = "Density"
+ )
It looks like most cell types have very few descendants, so let’s +zoom into the area below 500 to get a better look.
+ggplot(cl_df, aes(x = total_descendants)) +
+ geom_density(fill = "#FFCB05", alpha = 0.5) +
+ geom_vline(data = line_df,
+ mapping = aes(xintercept = total_descendants),
+ lty = 2) +
+ geom_text(
+ data = line_df,
+ mapping = aes(x = total_descendants, y = 0.6, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ labs(
+ x = "Number of descendants",
+ y = "Density"
+ ) +
+ xlim(c(0,500))## Warning: Removed 14 rows containing non-finite outside the scale range (`stat_density()`).## Warning: Removed 3 rows containing missing values or values outside the scale range (`geom_vline()`).## Warning: Removed 3 rows containing missing values or values outside the scale range (`geom_text()`).
Here we see a much larger range of values and that cell types become +more general as the number of descendants goes up. However, this +distribution alone is probably not helpful in determining a cutoff. The +next section we will look at this distribution specifically for cell +types present in our references, PanglaoDB and Blueprint encode.
+
+
+Latest common ancestor (LCA) between PanglaoDB and Blueprint +encode
+This section will look at identifying the latest common ancestor
+(LCA) between all possible combinations of terms from PanglaoDB (used
+for assigning cell types with CellAssign) and the
+BlueprintEncodeData reference from celldex
+(used for assigning cell types with SingleR). The LCA
+refers to the latest term in the cell ontology hierarchy that is common
+between two terms. I will use the ontoProc::findCommonAncestors()
+function to get the LCA for each combination.
Note that it is possible to have more than one LCA for a set of +terms. To start, I will keep all LCA terms found.
+For each LCA, I will again look at the total number of ancestors and +descendants and see if I can identify an appropriate cutoff. Ultimately, +I would like to see if we can use that cutoff to decide if we should +keep the LCA term as the consensus label or use “Unknown”.
+# first set up the graph from cl ont
+parent_terms <- cl_ont$parents
+cl_graph <- igraph::make_graph(rbind(unlist(parent_terms), rep(names(parent_terms), lengths(parent_terms))))# get a data frame with all combinations of panglao and blueprint terms
+# one row for each combination
+all_ref_df <- expand.grid(panglao_df$panglao_ontology,
+ blueprint_df$blueprint_ontology) |>
+ dplyr::rename(
+ panglao_ontology = "Var1",
+ blueprint_ontology = "Var2"
+ ) |>
+ # add in the human readable values for each ontology term
+ dplyr::left_join(blueprint_df, by = "blueprint_ontology") |>
+ dplyr::left_join(panglao_df, by = "panglao_ontology") |>
+ tidyr::drop_na() |>
+ dplyr::rowwise() |>
+ dplyr::mutate(
+ # least common shared ancestor
+ lca = list(rownames(ontoProc::findCommonAncestors(blueprint_ontology, panglao_ontology, g = cl_graph)))
+ )## Warning in dplyr::left_join(dplyr::left_join(dplyr::rename(expand.grid(panglao_df$panglao_ontology, : Detected an unexpected many-to-many relationship between `x` and `y`.
+## ℹ Row 49 of `x` matches multiple rows in `y`.
+## ℹ Row 99 of `y` matches multiple rows in `x`.
+## ℹ If a many-to-many relationship is expected, set `relationship = "many-to-many"` to silence this warning.## Warning: There were 23859 warnings in `dplyr::mutate()`.
+## The first warning was:
+## ℹ In argument: `lca = list(...)`.
+## ℹ In row 1.
+## Caused by warning in `dim()`:
+## ! The dim() method for DataFrameList objects is deprecated. Please use dims() on these objects instead.
+## ℹ Run `dplyr::last_dplyr_warnings()` to see the 23858 remaining warnings.lca_df <- all_ref_df |>
+ dplyr::mutate(
+ total_lca = length(lca), # max is three terms
+ lca = paste0(lca, collapse = ",") # make it easier to split the df
+ ) |>
+ # split each lca term into its own column
+ tidyr::separate(lca, into = c("lca_1", "lca_2", "lca_3"), sep = ",") |>
+ tidyr::pivot_longer(
+ cols = dplyr::starts_with("lca"),
+ names_to = "lca_number",
+ values_to = "lca"
+ ) |>
+ tidyr::drop_na() |>
+ dplyr::select(-lca_number) |>
+ # account for any cases where the ontology IDs are exact matches
+ # r complains about doing this earlier since the lca column holds lists until now
+ dplyr::mutate(lca = dplyr::if_else(blueprint_ontology == panglao_ontology, blueprint_ontology, lca)) |>
+ # join in information for each of the lca terms including name, number of ancestors and descendants
+ dplyr::left_join(cl_df, by = c("lca" = "cl_ontology")) ## Warning: Expected 3 pieces. Missing pieces filled with `NA` in 7967 rows [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
+## 18, 19, 20, ...].
+
+Distribution of ancestors and descendants
+ggplot(lca_df, aes(x = total_ancestors)) +
+ geom_density() +
+ geom_vline(data = ancestor_labels_df,
+ mapping = aes(xintercept = total_ancestors),
+ lty = 2) +
+ geom_text(
+ data = ancestor_labels_df,
+ mapping = aes(x = total_ancestors, y = 0.6, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ labs(
+ x = "Total number of ancestors",
+ y = "Density"
+ )
ggplot(lca_df, aes(x = total_descendants)) +
+ geom_density() +
+ geom_vline(data = line_df,
+ mapping = aes(xintercept = total_descendants),
+ lty = 2) +
+ geom_text(
+ data = line_df,
+ mapping = aes(x = total_descendants, y = 0.002, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ labs(
+ x = "Total number of descendants",
+ y = "Density"
+ )
Let’s zoom into the area below 1000, since we already know we would +want to exlude anything above that based on this plot.
+ggplot(lca_df, aes(x = total_descendants)) +
+ geom_density() +
+ geom_vline(data = line_df,
+ mapping = aes(xintercept = total_descendants),
+ lty = 2) +
+ geom_text(
+ data = line_df,
+ mapping = aes(x = total_descendants, y = 0.002, label = cl_annotation),
+ angle = 90,
+ vjust = -0.5
+ ) +
+ xlim(c(0, 1000)) +
+ labs(
+ x = "Total number of descendants",
+ y = "Density"
+ )## Warning: Removed 6856 rows containing non-finite outside the scale range (`stat_density()`).## Warning: Removed 1 row containing missing values or values outside the scale range (`geom_vline()`).## Warning: Removed 1 row containing missing values or values outside the scale range (`geom_text()`).
We can use the vertical lines for cells of interest to help us define +a potential cutoff based on the granularity we would like to see in our +consensus label. We want to be able to label things like T cell, but we +don’t want to label anything as lymphocyte as that’s probably not +helpful. I don’t see any obvious cutoffs that may be present in the +total number of ancestors, but the number of descendants is likely to be +informative. I think it might be a good idea to start by drawing a line +at the local maxima between the T cell and lymphocyte lines on the +number of descendants graph.
+
+
+Defining a cutoff for number of descendants
+First we will find the value for the first peak shown in the +distribution. This is likely to be a good cutoff for deciding which LCA +labels to keep.
+peak_idx <- splus2R::peaks(lca_df$total_descendants)
+cutoff <- lca_df$total_descendants[peak_idx] |>
+ min() # find the smallest peak and use that as the cutoff for number of descendants Below is the list of all consensus cell type labels that we will be +keeping if we were to just use this cutoff.
+celltypes_to_keep <- lca_df |>
+ dplyr::filter(total_descendants <= cutoff) |>
+ dplyr::pull(cl_annotation) |>
+ unique()
+
+celltypes_to_keep## [1] "myeloid leukocyte" "granulocyte" "neutrophil"
+## [4] "blood cell" "mononuclear phagocyte" "progenitor cell"
+## [7] "monocyte" "hematopoietic precursor cell" "T cell"
+## [10] "CD4-positive, alpha-beta T cell" "mature alpha-beta T cell" "mature T cell"
+## [13] "regulatory T cell" "memory T cell" "natural killer cell"
+## [16] "innate lymphoid cell" "B cell" "lymphocyte of B lineage"
+## [19] "mature B cell" "naive B cell" "memory B cell"
+## [22] "somatic stem cell" "stem cell" "hematopoietic stem cell"
+## [25] "bone cell" "macrophage" "erythroid lineage cell"
+## [28] "megakaryocyte" "endothelial cell" "lining cell"
+## [31] "dendritic cell" "eosinophil" "plasma cell"
+## [34] "chondrocyte" "stromal cell" "extracellular matrix secreting cell"
+## [37] "fibroblast" "smooth muscle cell" "muscle cell"
+## [40] "melanocyte" "cell of skeletal muscle" "ecto-epithelial cell"
+## [43] "keratinocyte" "squamous epithelial cell" "epidermal cell"
+## [46] "blood vessel endothelial cell" "microvascular endothelial cell" "adipocyte"
+## [49] "pericyte" "perivascular cell" "supporting cell"
+## [52] "astrocyte" "glial cell" "macroglial cell"
+## [55] "neuron associated cell" "mesangial cell"We can also look at all the cell types we are keeping and the total +number of descendants to see if there are any that may be we don’t want +to include because the term is too braod.
+# pull out the cell types and total descendants for cell types to keep
+plot_celltype_df <- lca_df |>
+ dplyr::filter(cl_annotation %in% celltypes_to_keep) |>
+ dplyr::select(cl_annotation, total_descendants) |>
+ unique()
+
+# bar chart showing total number of descendants for each cell type
+ggplot(plot_celltype_df, aes(x = reorder(cl_annotation, total_descendants), y = total_descendants)) +
+ geom_bar(stat = "identity") +
+ theme(
+ axis.text.x = element_text(angle = 90)
+ ) +
+ labs(
+ x = "cell type",
+ y = "Total descendants"
+ )
There are a few terms that I think might be more broad than we want
+like blood cell, bone cell,
+supporting cell, and lining cell. I’m on the
+fence about keeping myeloid leukocyte and
+progenitor cell. I think if we wanted to remove those terms
+we could move our cutoff to be the same number of descendants as
+T cell, since we do want to keep that.
One could also argue to remove stromal cell or
+extracellular matrix secreting cell.
Below are tables that look specifically at the combinations of cell +type annotations that resulted in some of the terms that I might +consider removing.
+
+
+Blood cell
+print_df <- lca_df |>
+ dplyr::select(blueprint_ontology, blueprint_annotation_main, blueprint_annotation_fine, panglao_ontology, panglao_annotation, total_lca, lca, cl_annotation)
+
+# blood cell
+print_df |>
+ dplyr::filter(cl_annotation == "blood cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000233 | +platelet | +2 | +CL:0000081 | +blood cell | +
| CL:0000232 | +Erythrocytes | +Erythrocytes | +CL:0000767 | +basophil | +2 | +CL:0000081 | +blood cell | +
| CL:0000232 | +Erythrocytes | +Erythrocytes | +CL:0000771 | +eosinophil | +2 | +CL:0000081 | +blood cell | +
| CL:0000232 | +Erythrocytes | +Erythrocytes | +CL:0000775 | +neutrophil | +2 | +CL:0000081 | +blood cell | +
| CL:0000232 | +Erythrocytes | +Erythrocytes | +CL:0000233 | +platelet | +2 | +CL:0000081 | +blood cell | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000233 | +platelet | +2 | +CL:0000081 | +blood cell | +
I think I’m in favor of not having a “blood cell” label, since I’m +not sure that it’s helpful. Also, if two different methods label +something a platelet and a neutrophil, then perhaps that label is +inaccurate and it’s really a tumor cell.
+
+
+Bone cell
+# bone cell
+print_df |>
+ dplyr::filter(cl_annotation == "bone cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000557 | +HSC | +GMP | +CL:0000092 | +osteoclast | +2 | +CL:0001035 | +bone cell | +
| CL:0000557 | +HSC | +GMP | +CL:0000137 | +osteocyte | +1 | +CL:0001035 | +bone cell | +
I think I would also remove bone cell, since hematopoietic stem cells +and osteoclasts seem pretty different to me.
+
+
+Myeloid leukocyte
+# myeloid leukocyte cell
+print_df |>
+ dplyr::filter(cl_annotation == "myeloid leukocyte")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000583 | +alveolar macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000235 | +macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000092 | +osteoclast | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000091 | +Kupffer cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000453 | +Langerhans cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000129 | +microglial cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000775 | +Neutrophils | +Neutrophils | +CL:0000874 | +splenic red pulp macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000583 | +alveolar macrophage | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000767 | +basophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000771 | +eosinophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000235 | +macrophage | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000775 | +neutrophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000092 | +osteoclast | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000091 | +Kupffer cell | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000453 | +Langerhans cell | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000129 | +microglial cell | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000874 | +splenic red pulp macrophage | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000767 | +basophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000771 | +eosinophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000775 | +neutrophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000092 | +osteoclast | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000453 | +Langerhans cell | +3 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000235 | +Macrophages | +Macrophages | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000767 | +basophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000771 | +eosinophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000775 | +neutrophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000092 | +osteoclast | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000453 | +Langerhans cell | +3 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000863 | +Macrophages | +Macrophages M1 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000767 | +basophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000771 | +eosinophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000775 | +neutrophil | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000092 | +osteoclast | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000453 | +Langerhans cell | +3 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000890 | +Macrophages | +Macrophages M2 | +CL:0000576 | +monocyte | +2 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000583 | +alveolar macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000235 | +macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000097 | +mast cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000092 | +osteoclast | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000091 | +Kupffer cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000453 | +Langerhans cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000129 | +microglial cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000889 | +myeloid suppressor cell | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000576 | +monocyte | +1 | +CL:0000766 | +myeloid leukocyte | +
| CL:0000771 | +Eosinophils | +Eosinophils | +CL:0000874 | +splenic red pulp macrophage | +1 | +CL:0000766 | +myeloid leukocyte | +
I’m torn on this one, because I do think it’s helpful to know if +something is of the myeloid lineage, but if we aren’t keeping lymphocyte +then I would argue we shouldn’t keep myeloid leukocyte. Noting that +after discussion we have decided to keep this one since T and B cells +are much easier to differentiate based on gene expression alone than +cells that are part of the myeloid lineage.
+
+
+Progenitor cell
+# progenitor cell
+print_df |>
+ dplyr::filter(cl_annotation == "progenitor cell") |>
+ head(n=15) # there's a lot of these so let's only print out some
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000765 | +erythroblast | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000062 | +osteoblast | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000158 | +club cell | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0000038 | +erythroid progenitor cell | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:4042021 | +neuronal-restricted precursor | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0002453 | +oligodendrocyte precursor cell | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000576 | +Monocytes | +Monocytes | +CL:0002351 | +progenitor cell of endocrine pancreas | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000765 | +erythroblast | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000576 | +monocyte | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000576 | +monocyte | +2 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000062 | +osteoblast | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000158 | +club cell | +1 | +CL:0011026 | +progenitor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000038 | +erythroid progenitor cell | +3 | +CL:0011026 | +progenitor cell | +
Same with progenitor cell, I do think it could be
+helpful to know that something may be a progenitor cell, but when you
+have a cell with the label for HSC and the label for cells like
+monocytes or osteoblasts, then maybe we are talking about a tumor cell
+instead. After discussion, we are going to remove progenitor cells.
Along those same lines, I think the below terms,
+lining cell and supporting cell, are too broad
+even though they have few descendants.
+
+Lining cell
+# lining cell
+print_df |>
+ dplyr::filter(cl_annotation == "lining cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000115 | +Endothelial cells | +Endothelial cells | +CL:0000077 | +mesothelial cell | +2 | +CL:0000213 | +lining cell | +
| CL:0000115 | +Endothelial cells | +Endothelial cells | +CL:0002481 | +peritubular myoid cell | +2 | +CL:0000213 | +lining cell | +
| CL:0000115 | +Endothelial cells | +Endothelial cells | +CL:0000216 | +Sertoli cell | +2 | +CL:0000213 | +lining cell | +
| CL:2000008 | +Endothelial cells | +mv Endothelial cells | +CL:0000077 | +mesothelial cell | +2 | +CL:0000213 | +lining cell | +
| CL:2000008 | +Endothelial cells | +mv Endothelial cells | +CL:0002481 | +peritubular myoid cell | +2 | +CL:0000213 | +lining cell | +
| CL:2000008 | +Endothelial cells | +mv Endothelial cells | +CL:0000216 | +Sertoli cell | +2 | +CL:0000213 | +lining cell | +
+
+Supporting cell
+# supporting cell
+print_df |>
+ dplyr::filter(cl_annotation == "supporting cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000669 | +Pericytes | +Pericytes | +CL:0000216 | +Sertoli cell | +2 | +CL:0000630 | +supporting cell | +
| CL:0000650 | +Mesangial cells | +Mesangial cells | +CL:0000216 | +Sertoli cell | +2 | +CL:0000630 | +supporting cell | +
+
+Discarded cell types
+We can also look at what cell type labels we are excluding when using +this cut off to see if there are any terms we might actually want to +keep instead.
+lca_df |>
+ dplyr::filter(total_descendants > cutoff) |>
+ dplyr::pull(cl_annotation) |>
+ unique()## [1] "leukocyte" "eukaryotic cell" "myeloid cell"
+## [4] "cell" "hematopoietic cell" "mononuclear cell"
+## [7] "stuff accumulating cell" "precursor cell" "phagocyte (sensu Vertebrata)"
+## [10] "defensive cell" "lymphocyte" "professional antigen presenting cell"
+## [13] "secretory cell" "connective tissue cell" "electrically responsive cell"
+## [16] "contractile cell" "epithelial cell" "neuron"
+## [19] "neural cell"The only terms in this list that I would be concerned about losing +are “neuron” and epithelial cells. Let’s look at those combinations.
+
+
+Neuron
+# neuron
+print_df |>
+ dplyr::filter(cl_annotation == "neuron")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000540 | +Neurons | +Neurons | +CL:0000109 | +adrenergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000108 | +cholinergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000166 | +chromaffin cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000700 | +dopaminergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0007011 | +enteric neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:1001509 | +glycinergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000099 | +interneuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000100 | +motor neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000165 | +neuroendocrine cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000540 | +neuron | +0 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0008025 | +noradrenergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000210 | +photoreceptor cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000740 | +retinal ganglion cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000850 | +serotonergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:4023169 | +trigeminal neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000695 | +Cajal-Retzius cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000617 | +GABAergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000679 | +glutamatergic neuron | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000121 | +Purkinje cell | +1 | +CL:0000540 | +neuron | +
| CL:0000540 | +Neurons | +Neurons | +CL:0000598 | +pyramidal neuron | +1 | +CL:0000540 | +neuron | +
It looks like there are a lot of types of neurons in the PanglaoDB +reference and only “neuron” as a term in Blueprint. Even though neuron +has ~ 500 descendants, I think we should keep these labels.
+
+
+Epithelial cell
+# epithelial cell
+print_df |>
+ dplyr::filter(cl_annotation == "epithelial cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000622 | +acinar cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:1000488 | +cholangiocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000166 | +chromaffin cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000584 | +enterocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000164 | +enteroendocrine cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000065 | +ependymal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000066 | +epithelial cell | +0 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000160 | +goblet cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000501 | +granulosa cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000182 | +hepatocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0005006 | +ionocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000312 | +keratinocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000077 | +mesothelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000185 | +myoepithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000165 | +neuroendocrine cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002167 | +olfactory epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000510 | +paneth cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000162 | +parietal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002481 | +peritubular myoid cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000652 | +pinealocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000653 | +podocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000209 | +taste receptor cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000731 | +urothelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002368 | +respiratory epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002370 | +respiratory goblet cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000171 | +pancreatic A cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000169 | +type B pancreatic cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000706 | +choroid plexus epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000158 | +club cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002250 | +intestinal crypt stem cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000173 | +pancreatic D cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002305 | +epithelial cell of distal tubule | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002079 | +pancreatic ductal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000504 | +enterochromaffin-like cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0005019 | +pancreatic epsilon cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002258 | +thyroid follicular cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002179 | +foveolar cell of stomach | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000696 | +PP cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000155 | +peptic cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002292 | +type I cell of carotid body | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0005010 | +renal intercalated cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:1000909 | +kidney loop of Henle epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002326 | +luminal epithelial cell of mammary gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002327 | +mammary gland epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000242 | +Merkel cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000682 | +M cell of gut | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002199 | +oxyphil cell of parathyroid gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000446 | +chief cell of parathyroid gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0005009 | +renal principal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002306 | +epithelial cell of proximal tubule | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002062 | +pulmonary alveolar type 1 cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002063 | +pulmonary alveolar type 2 cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:1001596 | +salivary gland glandular cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002140 | +acinar cell of sebaceous gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0000216 | +Sertoli cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002562 | +hair germinal matrix cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000066 | +Epithelial cells | +Epithelial cells | +CL:0002204 | +brush cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000622 | +acinar cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:1000488 | +cholangiocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000584 | +enterocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000164 | +enteroendocrine cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000066 | +epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000160 | +goblet cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000501 | +granulosa cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000182 | +hepatocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0005006 | +ionocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000185 | +myoepithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000510 | +paneth cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000162 | +parietal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000653 | +podocyte | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000209 | +taste receptor cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000731 | +urothelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002368 | +respiratory epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002370 | +respiratory goblet cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000171 | +pancreatic A cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000169 | +type B pancreatic cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000158 | +club cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002250 | +intestinal crypt stem cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000173 | +pancreatic D cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002305 | +epithelial cell of distal tubule | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002079 | +pancreatic ductal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000504 | +enterochromaffin-like cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0005019 | +pancreatic epsilon cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002258 | +thyroid follicular cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002179 | +foveolar cell of stomach | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000696 | +PP cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000155 | +peptic cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0005010 | +renal intercalated cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:1000909 | +kidney loop of Henle epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002326 | +luminal epithelial cell of mammary gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002327 | +mammary gland epithelial cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000682 | +M cell of gut | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002199 | +oxyphil cell of parathyroid gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0000446 | +chief cell of parathyroid gland | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0005009 | +renal principal cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002306 | +epithelial cell of proximal tubule | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:1001596 | +salivary gland glandular cell | +1 | +CL:0000066 | +epithelial cell | +
| CL:0000312 | +Keratinocytes | +Keratinocytes | +CL:0002204 | +brush cell | +1 | +CL:0000066 | +epithelial cell | +
The PanglaoDB cell types seem to be more specific than the ones
+present in Blueprint Encode, similar to the observation with neurons. We
+should keep epithelial cell in the cases where the Blueprint Encode
+annotation is Epithelial cells but not when it is
+Keratinocytes.
+
+Removing anything with more than 1 LCA
+One thing I noticed when looking at the labels that have less than +the cutoff is that most of them are from scenarios where we have +multiple LCAs. Maybe in the case where we have multiple LCAs we are +already too broad and we should just eliminate those matches from the +beginning. Here I’m looking at the total number of descendants for all +terms that show up because a term has multiple LCAs.
+lca_df |>
+ dplyr::filter(total_lca > 1) |>
+ dplyr::select(cl_annotation, total_descendants) |>
+ unique() |>
+ dplyr::arrange(total_descendants)
+
+
+
+
+| cl_annotation | +total_descendants | +
|---|---|
| bone cell | +38 | +
| blood cell | +41 | +
| perivascular cell | +41 | +
| stromal cell | +53 | +
| supporting cell | +61 | +
| hematopoietic precursor cell | +105 | +
| lining cell | +120 | +
| myeloid leukocyte | +165 | +
| progenitor cell | +165 | +
| mononuclear phagocyte | +169 | +
| phagocyte (sensu Vertebrata) | +175 | +
| contractile cell | +177 | +
| defensive cell | +199 | +
| professional antigen presenting cell | +212 | +
| connective tissue cell | +223 | +
| myeloid cell | +247 | +
| stuff accumulating cell | +266 | +
| precursor cell | +271 | +
| secretory cell | +457 | +
| mononuclear cell | +503 | +
| leukocyte | +540 | +
| electrically responsive cell | +673 | +
| hematopoietic cell | +684 | +
| eukaryotic cell | +2645 | +
It looks like most of these terms are pretty broad and are either +much higher than the cutoff or right around the cutoff with a few +exceptions. Things like “bone cell” and “supporting cell” have few +descendants, but I would still argue these are very broad terms and not +useful.
+I’m going to filter out any matches that show two LCA terms first and +then use the cutoff to define labels we would keep. I’ll also look to +see what cell types we lose when we add this extra filtering step to be +sure they are ones that we want to lose.
+# remove any combinations with more than one lca
+filtered_lca_df <- lca_df |>
+ dplyr::filter(total_lca < 2)
+
+# get a list of cell types to keep based on cutoff
+updated_celltypes <- filtered_lca_df |>
+ dplyr::filter(total_descendants <= cutoff) |>
+ dplyr::pull(cl_annotation) |>
+ unique()
+
+# which cell types are now missing from the list to keep
+setdiff(celltypes_to_keep, updated_celltypes)## [1] "blood cell" "hematopoietic precursor cell" "lining cell"
+## [4] "perivascular cell" "supporting cell"It looks like I am losing a few terms I already said were not +specific and then a few other terms, like “hematopoietic precursor cell” +and “perivascular cell”. I’ll look at both of those to confirm we would +not want them.
+
+
+Hematopoietic precursor cell
+print_df |>
+ dplyr::filter(cl_annotation == "hematopoietic precursor cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000050 | +HSC | +MEP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000050 | +HSC | +MEP | +CL:0000038 | +erythroid progenitor cell | +3 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000037 | +HSC | +HSC | +CL:0000038 | +erythroid progenitor cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000837 | +HSC | +MPP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000837 | +HSC | +MPP | +CL:0000038 | +erythroid progenitor cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000051 | +HSC | +CLP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000051 | +HSC | +CLP | +CL:0000038 | +erythroid progenitor cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000557 | +HSC | +GMP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000557 | +HSC | +GMP | +CL:0000038 | +erythroid progenitor cell | +3 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000049 | +HSC | +CMP | +CL:0000037 | +hematopoietic stem cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
| CL:0000049 | +HSC | +CMP | +CL:0000038 | +erythroid progenitor cell | +2 | +CL:0008001 | +hematopoietic precursor cell | +
It looks like here we should be keeping these matches because both +references have these labels as hematopoietic stem and progenitor cells. +I think in the context of pediatric cancer having this label would be +helpful, so maybe we shouldn’t remove all terms that have 2 LCAs.
+Let’s look at what the other LCA is for an example set.
+lca_df |>
+ dplyr::filter(panglao_ontology == "CL:0000037" & blueprint_ontology == "CL:0000050") |>
+ dplyr::select(blueprint_annotation_main, blueprint_annotation_fine, panglao_annotation, cl_annotation)
+
+
+
+| blueprint_annotation_main | +blueprint_annotation_fine | +panglao_annotation | +cl_annotation | +
|---|---|---|---|
| HSC | +MEP | +hematopoietic stem cell | +hematopoietic precursor cell | +
| HSC | +MEP | +hematopoietic stem cell | +progenitor cell | +
It looks like these terms have both
+hematopoietic precursor cell and
+progenitor cell as LCAs. Personally, I would keep the term
+for hematopoietic precursor cell because I think it’s more
+informative and specific to the type of progenitor cell.
+
+Perivascular cell
+print_df |>
+ dplyr::filter(cl_annotation == "perivascular cell")
+
+
+
+| blueprint_ontology | +blueprint_annotation_main | +blueprint_annotation_fine | +panglao_ontology | +panglao_annotation | +total_lca | +lca | +cl_annotation | +
|---|---|---|---|---|---|---|---|
| CL:0000669 | +Pericytes | +Pericytes | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000669 | +Pericytes | +Pericytes | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000669 | +Pericytes | +Pericytes | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000669 | +Pericytes | +Pericytes | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000650 | +Mesangial cells | +Mesangial cells | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000650 | +Mesangial cells | +Mesangial cells | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000650 | +Mesangial cells | +Mesangial cells | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
| CL:0000650 | +Mesangial cells | +Mesangial cells | +CL:0000359 | +vascular associated smooth muscle cell | +3 | +CL:4033054 | +perivascular cell | +
I would remove perivascular cell, since the cell type
+labels from PanglaoDB and Blueprint are pretty different from each
+other.
+
+Similarity index
+An alternative approach would be to calculate the similarity +index between each set of terms and define a cutoff for which set of +terms are similar. This is a value on a 0-1 scale where 0 indicates no +similarity and 1 indicates the terms are equal.
+Although this could provide a metric that we could use to define +similar cell types, we would still have to identify the label to use +which would most likely be the LCA. Even if the similarity index is +close to 1, if the LCA term is not informative then I don’t know that we +would want to use that.
+However, we could use this to finalize the actual pairs of terms that
+we trust. For example, if the LCA for a pair is T cell we
+can look at the similarity index to confirm that specific pair of terms
+has high similarity.
Below I’ll calculate the similarity index for each set of terms and +plot the distribution. Then we will look at the values for pairs that +have an LCA that pass the total descendants threshold we set to see if +those pairs have a higher similarity index.
+information_content <- ontologySimilarity::descendants_IC(cl_ont)
+
+# get similarity index for each set of terms
+si_df <- lca_df |>
+ dplyr::rowwise() |>
+ dplyr::mutate(
+ similarity_index = ontologySimilarity::get_sim_grid(ontology = cl_ont,
+ term_sets = list(panglao_ontology, blueprint_ontology)) |>
+ ontologySimilarity::get_sim()
+ )si_df <- si_df |>
+ dplyr::mutate(
+ lca_threshold = dplyr::if_else(total_descendants < cutoff, "PASS", "FAIL")
+ )
+
+ggplot(si_df, aes(x = similarity_index, fill = lca_threshold)) +
+ geom_density(bw = 0.05, alpha = 0.5) +
+ labs(
+ x = "Similarity index",
+ y = "Density"
+ )
This looks as I expected with most of the pairs that pass the total +descendants cutoff having a higher similarity index than those that do +not pass. There is still some overlap though so perhaps even if a set of +terms shares an LCA that passes the threshold, the actual terms being +compared may be further apart than we would like.
+Now let’s look at the similarity index for various LCA terms. Here +each LCA term is its own plot and the vertical lines are the similarity +index for each pair of terms that results in that LCA.
+celltypes_to_plot <- c("myeloid leukocyte", "T cell", "cell", "supporting cell", "B cell")
+
+celltypes_to_plot |>
+ purrr::map(\(celltype){
+ line_df <- si_df |>
+ dplyr::filter(cl_annotation == celltype) |>
+ dplyr::select(cl_annotation, similarity_index) |>
+ unique()
+
+ ggplot(si_df, aes(x = similarity_index)) +
+ geom_density() +
+ geom_vline(data = line_df,
+ mapping = aes(xintercept = similarity_index),
+ lty = 2) +
+ labs(
+ x = "Similarity index",
+ y = "Density",
+ title = celltype
+ )
+
+ })## [[1]]
##
+## [[2]]
##
+## [[3]]
##
+## [[4]]
##
+## [[5]]
It looks like terms that are more granular like T and B cell have +higher similarity index values than terms that are less granular which +is what we would expect. However, within terms like myeloid leukocyte +and even T cell we do see a range of values. We could dig deeper into +which pairs are resulting in which similarity index values if we wanted +to, but I think that might be a future direction if we feel like the +similarity index is something that could be useful.
+
+
+Conclusions
+Based on these findings, I think it might be best to create a +reference that has all possible pairs of labels between PanglaoDB and +Blueprint Encode and the resulting consensus label for those pairs. To +do this we could come up with a whitelist of LCA terms that we would be +comfortable including and all other cell types would be unknowns. I +would use the following criteria to come up with my whitelist:
+-
+
- Pairs should not have more than 1 LCA, with the exception of the +matches that have the label hematopoietic precursor cell. +
- The LCA should have equal to or less than 170 total +descendants. +
- We should include the term for
neuronand +epithelial celleven though they do not pass the threshold +for number of descendants. However,epithelial cellshould +only be included if the Blueprint Encode name is +Epithelial cellsand not +Keratinocytes.
+ - Terms that are too broad should be removed. This includes:
+
lining cell,blood cell, +progenitor cell,bone cell, and +supporting cell
+
Alternatively, rather than eliminate terms that are too broad we +could look at the similarity index for individual matches and decide on +a case by case basis if those should be allowed. Although I still think +having a term that is too broad, even if it’s a good match, is not super +informative.
+
+
+
+
+
+
+Session info
+sessionInfo()## R version 4.4.2 (2024-10-31)
+## Platform: aarch64-apple-darwin20
+## Running under: macOS Sequoia 15.2
+##
+## Matrix products: default
+## BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
+## LAPACK: /Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/lib/libRlapack.dylib; LAPACK version 3.12.0
+##
+## locale:
+## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
+##
+## time zone: America/Chicago
+## tzcode source: internal
+##
+## attached base packages:
+## [1] stats graphics grDevices datasets utils methods base
+##
+## other attached packages:
+## [1] ggplot2_3.5.1
+##
+## loaded via a namespace (and not attached):
+## [1] celldex_1.16.0 DBI_1.2.3 httr2_1.0.7 rlang_1.1.4
+## [5] magrittr_2.0.3 matrixStats_1.4.1 gypsum_1.2.0 compiler_4.4.2
+## [9] RSQLite_2.3.9 DelayedMatrixStats_1.28.0 png_0.1-8 vctrs_0.6.5
+## [13] pkgconfig_2.0.3 crayon_1.5.3 fastmap_1.2.0 dbplyr_2.5.0
+## [17] XVector_0.46.0 labeling_0.4.3 utf8_1.2.4 promises_1.3.2
+## [21] rmarkdown_2.29 tzdb_0.4.0 graph_1.84.0 UCSC.utils_1.2.0
+## [25] purrr_1.0.2 bit_4.5.0.1 xfun_0.49 zlibbioc_1.52.0
+## [29] cachem_1.1.0 splus2R_1.3-5 GenomeInfoDb_1.42.1 jsonlite_1.8.9
+## [33] blob_1.2.4 later_1.4.1 rhdf5filters_1.18.0 DelayedArray_0.32.0
+## [37] Rhdf5lib_1.28.0 parallel_4.4.2 R6_2.5.1 bslib_0.8.0
+## [41] reticulate_1.40.0 jquerylib_0.1.4 GenomicRanges_1.58.0 Rcpp_1.0.13-1
+## [45] SummarizedExperiment_1.36.0 knitr_1.49 readr_2.1.5 IRanges_2.40.0
+## [49] httpuv_1.6.15 Matrix_1.7-1 igraph_2.1.1 tidyselect_1.2.1
+## [53] abind_1.4-8 yaml_2.3.10 curl_6.0.1 ontologySimilarity_2.7
+## [57] lattice_0.22-6 tibble_3.2.1 shiny_1.9.1 Biobase_2.66.0
+## [61] withr_3.0.2 KEGGREST_1.46.0 evaluate_1.0.1 ontologyIndex_2.12
+## [65] BiocFileCache_2.14.0 alabaster.schemas_1.6.0 ExperimentHub_2.14.0 Biostrings_2.74.0
+## [69] pillar_1.9.0 BiocManager_1.30.25 filelock_1.0.3 MatrixGenerics_1.18.0
+## [73] DT_0.33 renv_1.0.11 stats4_4.4.2 generics_0.1.3
+## [77] vroom_1.6.5 rprojroot_2.0.4 BiocVersion_3.20.0 S4Vectors_0.44.0
+## [81] hms_1.1.3 sparseMatrixStats_1.18.0 munsell_0.5.1 scales_1.3.0
+## [85] alabaster.base_1.6.1 xtable_1.8-4 glue_1.8.0 alabaster.ranges_1.6.0
+## [89] alabaster.matrix_1.6.1 tools_4.4.2 ontologyPlot_1.7 AnnotationHub_3.14.0
+## [93] ontoProc_2.0.0 rhdf5_2.50.0 grid_4.4.2 tidyr_1.3.1
+## [97] AnnotationDbi_1.68.0 colorspace_2.1-1 GenomeInfoDbData_1.2.13 HDF5Array_1.34.0
+## [101] cli_3.6.3 rappdirs_0.3.3 fansi_1.0.6 S4Arrays_1.6.0
+## [105] dplyr_1.1.4 Rgraphviz_2.50.0 gtable_0.3.6 alabaster.se_1.6.0
+## [109] sass_0.4.9 digest_0.6.37 BiocGenerics_0.52.0 paintmap_1.0
+## [113] SparseArray_1.6.0 htmlwidgets_1.6.4 farver_2.1.2 memoise_2.0.1
+## [117] htmltools_0.5.8.1 lifecycle_1.0.4 httr_1.4.7 mime_0.12
+## [121] bit64_4.5.2