Mode of Inheritance classes

[mbrush]

Ticket #2 proposes a hierarchy of mode of inheritance terms for GENO that covers diversity of concepts in this area need for data ingest and interoperability. Here we discuss the ontological nature and definitions of these concepts so that we may create a clear and useful model of them for the larger community. The central question is how to frame them - as processes of inheritance? As qualities of inheritance processes? As qualities or dispositions of the trait/phenotype itself?

Some terminological clarification first. We define the following:

1. Trait = a characteristic that varies across organisms in a species or population (e.g. 'flower color')
2. Phenotype = the specific values of a trait (red, pink, white, mottled flowers)
3. Phenotype inheritance = the process by which a specific phenotype (e.g. pink flowers) is passed from one generation to the next based on the genetic interactions between alleles expressed at the responsible locus or loci (and contributions from the environment).

Approaches

1. Framing inheritance classes as processes (i.e subtypes of a root 'phenotype inheritance process') seems a relatively simple approach and avoids delving into the historically contentious realm of dependent continuants. It would require some relation of participation between a given trait/phenotype/disease and an inheritance class. For example:
   • 'baldness' participates_in or is output_of some 'X-linked inheritance'
   • diabetes participates_in or is output_of some 'polygenic inheritance'
2. Framing these classes as qualities of a trait or disease would mean some type of bearer_of relation between a given trait/phenotype/disease and an inheritance class. For example:

• 'baldness' has_quality some 'X-linked inheritance'
• diabetes has_quality some 'polygenic inheritance'

3. Framing these classes as qualities of an inheritance process would make the link from a given trait/phenotype/disease and an inheritance class trickier to define - as the term represents a quality that inheres not in the phenotype but in the process through which the phenotype is inherited.

[mellybelly]

I like the simplicity of (1). I feel like (2) is just odd, even though this is how it is often expressing in natural language, it really is not a quality of the phenotype itself. (3) is perhaps more ontologically correct, but very complicated to model.

If we go with (1), perhaps consider a different property, 'baldness' participates_in or is output_of some 'X-linked inheritance' seem funny, maybe it would be something more like 'baldness' inherits_over some 'X-linked inheritance' ??

[mbrush]

Thanks for the feedback Melissa. And yes, the participates_in example above was just meant to indicate some type of participation relation is needed if we frame inheritance types as processes. Something more meaningful/precise can be used. Your thoughts on all this @cmungall and @pnrobinson?

[cmungall]

process seems most extensible as its easier to slot more participants in.

How would alleles fit in, if at all? Might be useful to see some triples for Mendel's peas.

[mellybelly]

That should go in the paper!

On Sep 24, 2015, at 6:04 PM, Chris Mungall <notifications@github.commailto:notifications@github.com> wrote:

process seems most extensible as its easier to slot more participants in.

How would alleles fit in, if at all? Might be useful to see some triples for Mendel's peas.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/22#issuecomment-143095214.

[mbrush]

Interesting examples to inform modeling from ClinGen Modeling efforts here: http://datamodel.clinicalgenome.org/development/assertion/conceptual/condition_inheritance.html

[mbrush]

To consider:

Mode of Inheritance: Def = the disposition of a trait or condition (phenotype or disease) to manifest in an organism when a particular allelic complement is present in the host genome.

Autosomal dominant inheritance then would be defined as: A mode of inheritance whereby manifestation of a trait or condition requires inheritance of only one copy of a causative allele in an autosomal gene (i.e. the trait manifests in heterozygotes)