is_allele_of
A relation linking an instance of a variable feature (aka an allele) to a genomic location/locus it occupies. This is typically a gene locus, but a feature may be an allele of other types of named loci such as QTLs, or alleles of some unnamed locus of arbitrary size.
Note that the allele <fgf8a^ti282a> is not necessarily an instance of the danio rerio fgf8a gene class, given that we adopt the SO definition of genes as 'producing a functional product'. If the <fgf8a^ti282a> allele is nonfunctional or null, it is an allele_of the danio rerio fgf8a gene class, but not an instance (rdf:type) of this class. It would, however, bean instance of a 'danio rerio fgf8a gene allele' class - because being a 'gene allele' as defined in GENO requires only occupying the genomic position where for a gene, but not necessarily producing a functional product.
Domain = allele
Range = genomic locus (but in practice it is common to use a punned gene class IRI as the subject of this relation).
is_sequence_variant_of
To allow users to make important distinctions in discourse and modeling, GENO clearly separates the notions/levels of 'biological sequence', 'sequence feature', and 'sequence location' ('genomic locus' when found in a genome). This sets up an important terminological nuance when it comes to alleles, where we believe it correct to say that a particular genomic feature is an alleles_of some genomic locus (as opposed to an allele_of some sequence or some feature). This is typically a gene locus, but even insertions falling outside of genes are considered alleles_of the locus they alter (e.g. alleles of other types of named loci such as QTLs, or alleles of some unnamed locus of arbitrary size).
While conceptually it is most correct to say features are alleles_of some genomic locus, it is common practice to say that they are alleles of the class of feature defined to reside at that locus (typically a gene). Accordingly, we may write things like "fgf8a<ti282a> is an allele of the Danio rerio fgf8a gene", and we may create data where fgf8a<ti282a> is asserted as an allele_of the fgf8a gene class IRI. But here we mean more precisely that it is an allele of the locus at which the fgf8a gene resides. Allowing for this means that we dont have to create 'feature-based location/locus' terms mirroing all feature class terms already in exiistence (e.g. for every gene).
It is important to be clear that the location/locus that a feature is an allele_of is defined exclusively by its genomic position, and not on the sequence it may contains. This is particularly relevant when considering transgenic insertions. For example, this means that the insertion of the S. cerevisiae GAL4 gene sequence within the D. melanogaster Bx gene locus would create an allele of the D. melanogaster Bx gene, but not an allele of the S. cerevisiae GAL4 gene. The transgene that results from such an insertion, while expressing S. cerevisiae GAL4 gene sequence, is not an allele of this gene because it does not reside at the S. cerevisiae GAL4 locus.
This departs from how some databases use the term 'allele' - where transgenes expressing an exogenous gene are considered to be alleles of the exogenous genes they carry. For example, in the example above, Flybase describes the S. cerevisiae GAL4 transgene as an allele_of the S. cerevisiae GAL4 gene (and gives it the allele identifier FBal0040476). A GENO representation on the other hand would say that the S. cerevisiae GAL4 transgene derives_sequence_from the S. cerevisiae GAL4 gene, but is not an allele_of this gene. In a GENO model, FBal0040476 would be typed as a transgene insertion, but not considered an allele_of the Scer\GAL4 gene.
At the end of the day, it's just semantics, but worth clarifying given the ubiquity and variable use of the term 'allele'. The GENO model attempts to define and adhere to the principled notion of positionally-defined 'alleles', and functionally-defined 'transgenes'.
<fgf8a^ti282a> is_allele_of the 'danio rerio fgf8a' gene locus.
has_allele
has_affected_feature
genomic feature