Build general, upper-level slim

[dosumis]

The aim here is to tag a set of general cell types that would make sense to most biologists as grouping terms. Some judgement will be required in generating this - there is no perfect recipe, although we might gradually evolve some guidelines.

The terms chose should all be in the native cell (classification) branch, although the the immediate children of this term are mostly unsuitable (and any cells under here with no children after reasoning need review). Example of folding down to a level that is more suitable - neuron and muscle cell look appropriate but their parents do not.

Some terms may cross metazoans (neuron, muscle cell, epithelial cell), whereas other will be vertebrate specific (leukocyte). It may be useful to refer to text books on major vertebrate/mammalian cell types or use ChatGPT for inspiration

• [ ] Add upper_cell_type_slim AP with short definition - subProperty of http://www.geneontology.org/formats/oboInOwl#SubsetProperty (I'm open to better names!)
• [ ] Tag a set of high level cell types (see discussion above)
• [ ] Work with @ubyndr to test coverage under native cell - aiming for 80% coverage (excluding ancestors of slim terms)

[dosumis]

Chat GPT Experiment:

Neurons - specialized cells that transmit nerve impulses. Glial cells - support cells that provide structural and metabolic support for neurons. Epithelial cells - form the linings of internal and external surfaces of the body. Muscle cells - contractile cells responsible for movement. Adipocytes - fat cells that store energy. Hematopoietic stem cells - give rise to all types of blood cells. Endothelial cells - line blood vessels and regulate the exchange of materials between the blood and tissues. Fibroblasts - produce and maintain the extracellular matrix of connective tissue. Osteoblasts - bone-forming cells. Osteoclasts - bone-resorbing cells. Chondrocytes - cartilage-forming cells. Mast cells - release histamine and other substances in response to injury or inflammation. Macrophages - phagocytic cells that engulf and destroy pathogens and cellular debris. T cells - immune cells that coordinate the immune response. B cells - immune cells that produce antibodies. Natural killer cells - immune cells that kill infected or cancerous cells. Dendritic cells - immune cells that present antigens to T cells. Hepatocytes - liver cells that perform a variety of metabolic functions. Pancreatic beta cells - produce insulin. Pancreatic alpha cells - produce glucagon. Kidney cells - perform a variety of functions related to waste removal and fluid balance. Sertoli cells - support cells in the testes that help to produce sperm. Leydig cells - produce testosterone in the testes. Granulosa cells - support cells in the ovaries that help to produce eggs. Theca cells - produce hormones in the ovaries. Keratinocytes - produce the protein keratin, which forms the basis of hair, nails, and the outer layer of skin. Melanocytes - produce the pigment melanin, which gives color to the skin, hair, and eyes. Langerhans cells - immune cells in the skin that help to protect against infection. Smooth muscle cells - contractile cells found in the walls of internal organs, blood vessels, and airways. Cardiomyocytes - specialized muscle cells that make up the heart.

[addiehl]

Interesting result, but I wonder if 30 is too low to catch all the major cell types. Nice representation of major immune cells in this list though.

[aleixpuigb]

Proposed definition for upper_cell_type_slim: "a subset of general classes in the cell ontology."

---

High level terms (some are very general (tagged with *) and will be assessed later):

neuron glial cell epithelial cell secretory cell *muscle cell adipocyte endothelial cell fibroblast osteoblast osteoclast chondrocyte mast cell macrophage T cell B cell hematopoietic stem cell natural killer cell dendritic cell hepatocyte enteroendocrine cell neuroendocrine cell kidney cell supporting cell smooth muscle cell cardiac muscle cell skeletal muscle cell striated muscle cell male germ cell female germ cell seromucus secreting cell pancreatic endocrine cell keratinocyte melanocyte retinal cell sensory receptor cell extraembryonic cell myeloid cell glandular epithelial cell

[emquardokus]

@aleixpuigb @dosumis To be honest, I think it makes the most sense to start by looking at the current hierarchy of CL at tree level and it becomes really obvious what doesn't make sense, then figure out at what level you want to hit and the hard part is to figure out how computationally you can retrieve the same list! It's most natural to biologist to think in terms of histology to distinguish tissues and the cells that you can observe from various antibody based imaging methods. Some thought of using the embryonic layer origins as a starting place: The three germ layers are the endoderm, the ectoderm, and the mesoderm. (This has been discussed before and some loose categorization based on this in CL). Upper cell types should match the tissue / histology level types: 4 types high level: 1) connective 2) epithelial, 3) muscle 4) nervous Each of those can be broken down further: 7 Connective tissues: 1) cartilage 2) bone (osseus), 3) blood, 4) adipose (fat), 5) loose connective tissue 6) dense fibrous connectve tissue 7) elastic connective tissue 3 muscle tissue types: 1) striated = a) skeletal (voluntary) or b) cardiac) , c smooth muscle 6 Epithelial tissue types: The number of cell layers and cell types together give rise to 6 different types of epithelial tissue. Simple squamous epithelia. Simple cuboidal epithelia. Simple columnar epithelia. Stratified squamous epithelia. Stratified cuboidal epithelia. Stratified columnar epithelia. Plus endothelium and mesothelium

Also I was looking at other categories of cell types and wonder if some are redundant or not useful. For instance categorizing on number of nuclei is okay but by far the majority of all cells are single nucleate but all single nucleate cell types are not listed under it. The real minority cells are multi nucleate. All of this is mostly categorizing white blood cell types or could be muscle cell types. Not useful for most other cell types. Similar vein is categorizing on haploid, diploid or polyploid. I think these are covered by the nature of the cell type itself and the vast majority are diploid. Yet you don’t see all diploid cells under diploid. For epithelial tissue types there are more actually that you learn in histology course like transitional epithelium (found in urinary bladder. I just pulled done quick stuff out as examples.

[dosumis]

This seems reasonable as a starting point for coverage:

cartilage 2) bone (osseus), 3) blood, 4) adipose (fat), 5) loose connective tissue 6) dense fibrous connectve tissue 7) elastic connective tissue 3 muscle tissue types: striated = a) skeletal (voluntary) or b) cardiac) , c smooth muscle

• eptihelial cell
• immune cells
• but I'm sure this won't cover enough of the ontology - so maybe start here and extend with other grouping types until coverage is sufficient.

[dosumis]

@bvarner-ebi @aleixpuigb - Can we make it a priority to get something released here. Needed for Pandasaurus. We can iterate on it to improve once an initial version is out.