Gene and shared-gene associations

Most of Evagene's correlation graph is hand-curated, but the genetic edges are not — Evagene reads its marker catalogue and generates them automatically. Two kinds of edge fall out: a gene linked to the diseases, traits, and allergies it is associated with, and the entities that share a gene with one another. This is genotype–phenotype exploration for teaching and research, not risk prediction.

| 7 min read

Educational reference data. "Shares a gene" is a statement about the catalogue, not about a person. These edges teach which catalogue entities are linked to the same gene so a learner can explore pleiotropy and shared genetic architecture. They produce no risk number, run no inheritance analysis, and assert nothing about any individual.

Two auto-derived edge types

Where the curated edges are written by hand with their own notes, the genetic edges are generated on the fly from the marker catalogue and tagged source: marker-derived. There are two:

  • genetic_association — joins a gene / marker to each disease, trait, or allergy it is linked to in the marker catalogue. This is the gene-to-phenotype edge.
  • shared_gene — joins any two catalogue entities that are linked to the same gene / marker. This is the entity-to-entity edge that surfaces "these share a gene".

Because they are regenerated from the marker catalogue at start-up, they stay in step with it automatically: add a gene–disease link to the catalogue and the corresponding edges appear without anyone hand-listing them. The curated genetic associations should therefore not be authored by hand — the marker catalogue is the single source for them.

The gene-to-phenotype edge in practice

A single gene is often linked to several entities, and not always of the same kind. A few examples drawn straight from the marker catalogue:

GeneLinked entities (genetic_association)
BRCA2Breast cancer · ovarian cancer · pancreatic cancer · hereditary breast and ovarian cancer
PTENBreast cancer · endometrial cancer · thyroid cancer · Cowden syndrome
GBAGaucher disease · Parkinson's disease (familial)
MITFMelanoma (disease) · eye colour (trait)

The MITF row shows the cross-kind case: the same gene links a disease and a trait, so the graph offers an educational bridge between concepts a learner might otherwise treat as unrelated. The GBA row is a classic teaching link between a recessive lysosomal storage disease and a neurodegenerative association — exactly the kind of connection a concept graph is built to surface.

The shared-gene edge

Wherever two entities are linked to the same gene, a shared_gene edge connects them directly. Because BRCA2 links breast, ovarian, and pancreatic cancer (and the HBOC entry), each of those pairs is joined by a shared-gene edge. So with breast cancer recorded on an individual, the Related concepts panel can surface ovarian and pancreatic cancer under "shares a gene (BRCA2)", letting a learner explore the shared architecture.

"Shares a gene" is not "carries the same risk". A shared-gene edge says two conditions are linked to the same gene in the catalogue. It does not say they share a mechanism, a penetrance, or a recurrence risk, and it makes no statement about any person. For inheritance probabilities and published risk models, Evagene's separate Mendelian inheritance and hereditary cancer risk tools are the right place — and they are themselves explicitly educational implementations.

How it appears in the editor

With a marker or a gene-linked condition recorded on an individual, the Related concepts panel groups genetic suggestions under the recorded item — for example "Because BRCA2 is recorded:" or "shares a gene with Breast cancer:". Ticking a suggestion records it on the individual; the graph never records anything by inference. The "explore any concept" search lets a learner pull up everything linked to a gene without recording it on anyone, which makes it a quick way to teach pleiotropy and the one-gene-many-phenotypes idea.

Boundaries

  • Genetic edges are derived from catalogue links, not from any individual's genotype, and they assert nothing about any person.
  • A shared_gene edge is a catalogue relationship, not a claim of shared risk, shared mechanism, or shared inheritance.
  • No risk number is produced and no inheritance analysis is run from these edges.
  • Evagene is an academic, research, and educational pedigree modelling platform — not a medical device, not clinical decision support, and not a diagnostic or screening tool.

Further reading

Explore the Related Concepts panel

No credit card, no enterprise sales cycle — free access during Alpha for clinicians, researchers, educators, and students.

Join the Alpha Waiting List