Related Concepts Explorer — Evagene's Educational Correlation Graph

Educational reference data. The correlation graph sits in the same family as Evagene's disease, trait, allergy, and clinical-test catalogues. It describes which concepts tend to be associated with one another. It does not diagnose, it asserts nothing about any individual, and it produces no risk number. A concept is recorded on a person only when a user records it — never by inference.

What the correlation graph is

A clinician or educator working in Evagene records concepts on individuals in a pedigree: a disease on an affected relative, a recorded clinical-test result, an allergy, a trait, a genetic marker. The correlation graph is a catalogue of educational associations between exactly those kinds of entities. Where two catalogue entries are recognised as associated — a low ferritin and iron-deficiency anaemia, two conditions that share a gene, a finding that commonly accompanies a condition — the graph records an edge between them with a short, neutral explanation.

It is the same shape of artefact as the disease catalogue or the help-catalogue guides: curated reference material a learner can browse. It teaches associations so they can be explored; it never states that a subject has a condition.

How an association is modelled

Each association is an edge joining two endpoints. An endpoint has a kind — one of five: disease, trait, clinical-test result, allergy, or marker — and names a real entry in the corresponding catalogue. Every edge also carries:

relationship — a term from a fixed, eight-term vocabulary (below), so associations stay consistent and mean the same thing everywhere.
note — a short, neutral educational explanation of the association.
strength — optionally strong, moderate, or weak.
directional — whether the association reads naturally in one direction (a finding pointing to a condition) rather than both.
source — curated for hand-authored edges, marker-derived for edges Evagene generates automatically from the marker catalogue.

The relationship vocabulary

Relationship	Meaning
`genetic_association`	A gene / marker is associated with a disease, trait, or allergy.
`shared_gene`	Two entities linked to the same gene / marker.
`biomarker`	A clinical-test result that acts as a laboratory marker associated with a condition (e.g. a raised fasting glucose and type 2 diabetes).
`associated_finding`	A test result or observation that commonly accompanies a condition (e.g. a low mean cell haemoglobin alongside iron-deficiency anaemia).
`associated_condition`	Two conditions recognised as associated with one another.
`shares_features`	Two entities that share overlapping features.
`sequela`	A consequence that can follow on from a condition.
`risk_association`	A factor associated with a raised likelihood of a condition.

Two sources: curated and marker-derived

The graph is assembled from two sources at start-up, giving both the considered associations a curator wants to teach and the broader "these share a gene" connections that fall out of the marker catalogue:

Curated edges — over 1,100 hand-authored associations, each written with its own educational note and relationship term. These are the considered connections: biomarkers and findings, associated conditions, shared features, sequelae, and risk associations.
Auto-derived genetic edges — Evagene reads the marker catalogue and adds two kinds of edge on the fly, both tagged marker-derived: a genetic_association edge between a gene / marker and each disease, trait, or allergy it is linked to, and a shared_gene edge between any two entities linked to the same marker.

Status-qualified clinical-test edges

A clinical-test result means different things depending on whether it is low or high, so a clinical-test endpoint can be qualified by a status. The graph then surfaces different associations for each direction. The headline example is ferritin:

Ferritin — low

Associated with iron-deficiency anaemia and restless legs syndrome.

Ferritin — high

Associated with hereditary haemochromatosis.

When an individual has a ferritin value recorded that comes out low, the panel offers the low-ferritin associations; when it comes out high, it offers the high-ferritin association instead. A test that is recorded but not marked low or high surfaces only the associations that do not depend on direction — never the direction-specific ones. See the biomarker and condition associations page for more on clinical-test edges.

Using it in the editor

With an individual selected, open the Related concepts panel from the dock button between Genetics and Risk analysis. It lists concepts associated with the items already recorded on that person — diseases, traits, clinical-test results (including their low/high status), allergies, and markers — grouped per recorded item, for example "Because Ferritin (low) is recorded:". Items already on the individual show ticked; ticking a suggestion adds it to the person and unticking removes it. An "explore any concept" search lets you browse associations for any catalogue entity without recording it on anyone. The educational-use banner is shown by default and can be suppressed per user from the Account tab.

Because the panel surfaces a concept as "already recorded" only when it genuinely is, and nothing is ticked by inference, the learner always does the asserting. The graph offers the associations; the user decides what, if anything, to record.

API and MCP access

The same data is available programmatically. The open GET/POST /api/correlations* endpoints return the full graph or the concepts related to one or more anchors — and, being static reference data that produces no risk number, they are not gated by the intended-use acknowledgement. For AI agents, the MCP related_concepts tool returns related concepts for a given anchor and always carries the educational-use disclaimer, so an external agent receives the framing alongside the data. See the Evagene MCP server and platform and integrations pages.

What the correlation graph does not do

It does not diagnose, and it does not assert that any person has any condition.
It produces no risk number and runs no inheritance analysis. For those, Evagene's separate Mendelian inheritance and hereditary cancer risk tools are explicitly educational and research implementations of published models.
Nothing is pre-selected by inference; a concept appears as "recorded" only when a user has recorded it.
Notes are neutral and educational. They describe an association; they do not tell anyone what to do, and they carry no screening, referral, or treatment instruction.
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.

Related Concepts: Evagene's educational correlation graph