Main Page

From Cell Types
Jump to navigation Jump to search

Cell types

What is a cell type?

This wikibase is a shot at formalizing meanings for cell types and cell type descriptors. There are, of course, well structured efforts with similar goals, such as the [Cell Ontology https://github.com/obophenotype/cell-ontology]. This is a bit more exploratory, and admittedly less structured.

Formal criteria

The notion is that anything that any class of cells that fits the 3 following criteria is formally a cell type:

  • The instances are rigorously defined (i.e. you have guidelines for telling wheter a cell is or isn't an instance of that class)
  • The class' instances are present in a specific taxon
  • The class is considered useful by any research project.

Some subdefinitions help to organize it.

Note that these are not considered axioms, as the second (* The class' instances are present in a specific taxon) can be partially inferred from the first, as this might be needed for a rigorous definition.

Subdefinitions

Cell archetype

A cell type defined for more than one species, e.g. "neuron")

Cell type strictu sensu

A cell type defined for one taxon (e.g. "human neuron")

Cell infratype

A cell type defined for a taxon below the species level (e.g. "C57BL/6J mice neuron")

Technotype

A cell type that rigorously corresponds to what is being measured in a given experiment ("ex: CD4+ CD8- FACS-sorted C57BL/6J mice cell")


Core items and properties

Items

Which are all instances of:


Properties

Basic classification properties

Descriptor properties

These are properties that all items for instances of cell types should have

  • has scope - the taxon or taxons in the scope for this conceptualization of cell type (e.g. "human neuron" has scope "Homo sapiens")
  • natural language definition - a natural language definition of this cell type. (e.g. "human neuron" has natural language definition "cells capable of transmitting electrical impulses).
    • Note that the natural language definition will likely have some ambiguity, as this is a feature of natural languages. Ideally, the description should be described by dedicated properties, allowing computable definitions, but this is an arduous, at least decade-long task. The scope doesn't need to be stated again in the natural language definition, as it is already parsed my has scope.

Model items