Consistent indication of organism specificity of terms in CL

addiehl commented 4 years ago

This issue address Proposal 1 of Bjoern Peters @bpeters42 from https://docs.google.com/document/d/1bSpiZc5p9VNCZ83I4VUEwOVGoFTZ3cpXZKOSRAaGBZc/edit?usp=sharing

proposed hierarchy based on existing terms in CL and incorporating Proposal 1

cell --native cell ----eukaryotic cell [CL:0000255] ------cell from multicellular organism [new] --------animal cell [CL:0000548] ------cell from obligate unicellular Eukaryote [new] ----unicellular organism [new] ------prokaryotic cell [CL:0000520] ------cell from obligate unicellular Eukaryote [new]

We would want to create appropriate logical definitions for all the above cell types where possible. These changes are perhaps non-controversial.

Below are additional grouping classes proposed by Bjoern and some example classes that would be reasoned to be subclasses of them.

--native cell ----cell with certain part [new] ------nucleate cell [CL:0002242] ------anucleated cell [CL:0000225] ----cell with certain function [new] ------photosynthetic cell [CL:0000628] ------cell with specialized function in organism [new] --------germ cell [CL:0000586] --------absorptive cell [CL:0000003] ----eukaryotic cell [CL:0000255] ------cell from multicellular organism [new] --------cell with characteristic anatomical location [new] ----------epithelial cell [CL:0000066] ----------cardiocyte [C:0002494] --------cell with specialized function in organism [new] ----------germ cell [CL:0000586] ----------absorptive cell [CL:0000003] --------cell with characteristic differentiation lineage [new] ----------hematopoietic cell [CL:0000988]

In order for this schema to work well, so that we don't fall into the faults of the original Cell Ontology and its arbitrary hierarchies, we will to create appropriate logical definitions for all the high level cell types that enable the creation of an appropriate poly hierarchy of granular cell types that satisfy the differentia for multiple high level grouping classes.

And there are existing classes such as 'animal cell' and 'somatic cell' that need to be placed appropriately in this new schema.

dosumis commented 4 years ago

metaclasses like 'cell with certain part' are useful for organising the tree for browsing purposes but may be hard to maintain.

Can these classes be maintained using design patterns or will they be manually maintained ? If we try to use logical definitions, almost everything will end up under everything, - every cell has some 'certain' part, most have characteristic lineage, anatomical location etc.

bpeters42 commented 4 years ago

this is solely about design patterns, as spelled out in my original document
yes, parentage to these 'organizing classes' like 'cell with part' should never be asserted, but only inferred based on the design patterns. And yes, ideally each cell gets inferred in multiple places, essentially giving different browsing hierarchies.

On Wed, Mar 25, 2020 at 8:29 AM David Osumi-Sutherland < notifications@github.com> wrote:

metaclasses like 'cell with certain part' are useful for organising the tree for browsing purposes but may be hard to maintain.

Can these classes be maintained using design patterns or will they be manually maintained ? If we try to use logical definitions, almost everything will end up under everything, - every cell has some 'certain' part, most have characteristic lineage, anatomical location etc.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/obophenotype/cell-ontology/issues/644#issuecomment-603906937, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADJX2IULXFWMG64YA6CSUF3RJIPOFANCNFSM4LTQLN5A .

-- Bjoern Peters Professor La Jolla Institute for Allergy and Immunology 9420 Athena Circle La Jolla, CA 92037, USA Tel: 858/752-6914 Fax: 858/752-6987 http://www.liai.org/pages/faculty-peters

obophenotype / cell-ontology

Consistent indication of organism specificity of terms in CL #644