Expected value, Value syntax

[lschriml]

Current term details Specifications to be made consistent across MIxS

• When an ontology term is specified, to be used, in the definition.

• Issue, some terms have for Expected value 'text or', other only specify the ontology

  • I would rather we do not allow 'text' as an option, rather than using the specified ontology,

Expected value Value syntax Example text or OBI {text}|{termLabel} {[termID]}

update to: Example OBI term, OBI ID time series design [EFO:EFO_0001779]

Note: the Value syntax varies across terms that use and ontology. --> I propose we unify this format to:

{text} {termLabel}:{[termID]}

Cheers, Lynn

as seen for terms:

experimental_factor samp_mat_process growth facility

[mia-lgo]

Hello @lschriml ,

I joined the gensc-cig group back in November, so I might be missing some context here. Please correct me if I'm wrong.

Here, my humble opinion for the terms with expected value 'text or' :

In cases such as 'experimental_factor' having 'text or EFO and/or OBI' allows the alternative to specifying the experiment even when there is no single EFO or OBI term (even in combination) that applies.

• I am mentioning this because we've run into a similar situation with a submission we received in which adding the 'experimental_factor' field using 'text' proved to be the only way to understand the experimental setup and the relationship between the MIxS terms used.

Leaving the option 'text or EFO and/or OBI' for non-mandatory MIxS fields (such as experimental_factor, samp_mat_process, etc.) allows (both researchers and curators) to provide details/context as 'text' when there is no appropriated ontology term(s).

Perhaps specifying that an ontology term is preferred over 'text' might help to encourage the use of EFO or OBI.

I hope I understood this right.

[mia-lgo]

Hi @lschriml let me know what are your thoughts.

[lschriml]

@mia-lgo - agreed, will leave the fields as is. In the definition, we do specify preference for using ontology terms.

Cheers, Lynn