uPheno integration issues

[anna-anagnostop]

While searching for cellular phenotypes in uPheno I have come across a number of integration issues, including:

1. nucleus organization phenotype UPHENO:0083146

• equivalent to 'has part' some ('process quality' and ('characteristic of' some 'nucleus organization') and ('has modifier' some abnormal))
• EQ uses the GO term 'nucleus organization' GO:0006997 (which has related synonym = nuclear morphology)
• I would have expected ‘nucleus organization phenotype’ (currently only under 'organelle organization phenotype' UPHENO:0049700) to be an additional child of ‘nucleus phenotype’ UPHENO:0082548 (def: changed nucleus) but it is not
• Note the EQ for ‘nucleus phenotype’ uses 'has part' some (quality and ('characteristic of part of' some nucleus) and ('has modifier' some abnormal))
• As a side note, the definition and term label for ‘nucleus phenotype’ are rather ambiguous; consider changing the term label to 'cell nucleus phenotype' to clearly differentiate it from a brain nucleus

2. abnormal eosinophil morphology MP:0005061

• equivalent to 'has part' some (morphology and ('characteristic of' some eosinophil) and ('has modifier' some abnormal))
• I would have expected MP:0005061 to be an additional child of 'eosinophil phenotype' UPHENO:0086175 but it is not
• I only find the HPO term 'Abnormal eosinophil morphology' HP:0001879 under 'eosinophil phenotype' as HP:0001879 and UPHENO:0086175 share the same EQ 'has part' some (quality and ('characteristic of part of' some eosinophil) and ('has modifier' some abnormal))
• Moreover, 'morphology of eosinophil phenotype' UPHENO:0079609 has no child terms, yet its EQ is identical to the EQ for MP:0005061

3. decreased kidney apoptosis MP:0011371

• equivalent to 'has part' some ('decreased rate' and ('characteristic of part of' some ('apoptotic process' and ('occurs in' some kidney))) and ('has modifier' some abnormal)) -- based on the abnormallyDecreasedRateOfBiologicalProcessInLocation pattern but uses 'characteristic of part of' instead of 'characteristic of'
• I would have expected MP:0011371 to be a child of 'decreased apoptotic process' UPHENO_0005644 but it is not
• EQ for 'decreased apoptotic process' uses 'has part' some ('decreased rate' and ('characteristic of' some 'apoptotic process') and ('has modifier' some abnormal)) -- based on the abnormallyDecreasedRateOfBiologicalProcess pattern
• It would seem appropriate to have any terms using the inLocation pattern placed under 'decreased apoptotic process'
• The only terms I find under 'decreased apoptotic process' are HPO terms like 'Decreased lymphocyte apoptosis' HP:0002731 whose EQ uses 'lymphocyte apoptotic process' GO:0070227 where location (i.e. lymphocyte) is already built into the GO term

• MP terms like 'decreased T cell apoptosis' MP:0006414 and 'decreased B cell apoptosis' MP:0008783 do not appear under 'decreased apoptotic process' but have EQs that use 'characteristic of part of' and the built-in location GO terms for 'T cell apoptotic process' and 'B cell apoptotic process' respectively
• As a side note, 'apoptotic anatomical entity' UPHENO:0076955 has no child terms, yet ZP has >200 terms that use the 'abnormallyApoptoticAnatomicalEntity' pattern

[ValWood]

A question. What does "nucleus phenotype" actually mean?

If it means any phenotype that is observed in the nucleus, it would be tricky. So does this term really mean "nucleus organization phenotype"? or "nucleus morphology phenotype (size and shape)"

[sbello]

@ValWood nucleus phenotype is meant to be the grouping class that collects all morphological and physiological abnormalities in the nucleus. The removal of the word abnormal from the uPheno labels without replacing it with another word with a similar if less negative connotations is I think contributing to the confusion. Both organization and morphology terms should fall under this term. If this isn't a useful grouping class than we should also discuss that. We can add that to the potential workshop topics.

[ValWood]

got it! In FYPO, due to a historic design decision, we keep abnormal morphology and abnormal process phenotypes distinct. Despite being (probably) ontologically correct I think that searches would benefit from such a grouping term (because often the genes annotated to such terms obviously often overlap). In fact, in some cases, it is difficult to distinguish, and really only depends on the method of observation (for example, we were recently unable to justify abnormal cytoskeleton organization from abnormal cytoskeleton morphology). I'm still considering what to do here.

[sbello]

We've done a similar thing in the MP and have a similar issue, are they talking about the process or the outcome of the process is frequently tricky to decide.