A study reports that the gene {gene} is associated with:

```
"{phenotype}"
```

Below are traits already linked to {gene} in a public database. They are the 20
nearest entries by embedding similarity, selected from hundreds of traits.
Most are expected to be near-misses rather than true matches.

Your task is to determine whether each existing gene-trait association provides
meaningful prior evidence for the reported gene-trait association.

For each entry, classify its relationship to the reported trait using exactly
one of the following tags.

EXACT
The same trait: synonyms, abbreviations, spelling variants, or trivially
different measurements of essentially the same quantity.

```
Example:
  Reported: "systolic blood pressure"
  Existing: "SBP"
  -> EXACT
```

PARENT_FIRST
The existing trait contains the reported trait and is close enough that its
association provides meaningful prior evidence.

```
Example:
  Reported: "stage III breast cancer"
  Existing: "breast cancer"
  -> PARENT_FIRST
```

PARENT_DISTANT
The existing trait contains the reported trait, but is broad enough to
include substantially different phenotypes, so it is not meaningful prior
evidence.

```
Example:
  Reported: "stage III breast cancer"
  Existing: "cancer"
  -> PARENT_DISTANT
```

CHILD_FIRST
The existing trait is contained within the reported trait and is close
enough that its association provides meaningful prior evidence.

```
Example:
  Reported: "breast cancer"
  Existing: "stage III breast cancer"
  -> CHILD_FIRST
```

CHILD_DISTANT
The existing trait is contained within the reported trait, but is much more
specific, so it is not meaningful prior evidence for the broader trait.

```
Example:
  Reported: "cancer"
  Existing: "stage III breast cancer"
  -> CHILD_DISTANT
```

PARENT and CHILD require meaningful phenotypic containment. Do not infer them
merely from shared ontology membership, organ system, correlation, causal
influence, or mathematical dependence.

FIRST versus DISTANT is determined by how informative the existing association
is for the reported trait, not simply by the number of ontology levels between
them.

CLOSE_VERY
Different traits that capture nearly the same underlying biological
quantity, such that an association with one is very strong prior evidence
for association with the other.

```
Example:
  Reported: "hemoglobin concentration"
  Existing: "hematocrit"
  -> CLOSE_VERY
```

CLOSE_MODERATE
Different traits linked by one concrete and specific biological or
phenotypic axis, such that an association with one provides meaningful prior
evidence for association with the other.

```
Example:
  Reported: "red blood cell count"
  Existing: "hemoglobin concentration"
  -> CLOSE_MODERATE

Different units or measurement types do not prevent CLOSE_MODERATE. However,
sharing only an organ, broad field, sample type, general risk factor, or
broad correlation is not enough.
```

CLOSE_VAGUE
A specific biological or epidemiological relationship exists, but an
association with one is too weak or unreliable to provide meaningful prior
evidence for the other.

```
Example:
  Reported: "HDL cholesterol"
  Existing: "LDL cholesterol"
  -> CLOSE_VAGUE

Mere membership in the same broad biological field is DIFFERENT, not
CLOSE_VAGUE.
```

If one trait is primarily a manifestation, consequence, complication, or marker
of the other, use SYMPTOM rather than CLOSE.

SYMPTOM_SPECIFIC
One trait is defining, near-constant, or directly used to diagnose the
other. The relationship provides meaningful prior evidence.

```
Example:
  Reported: "A1C >=6.5%"
  Existing: "diabetes mellitus"
  -> SYMPTOM_SPECIFIC
```

SYMPTOM_MODERATE
One trait is a common and characteristic manifestation or consequence of the
other, though not defining or highly specific. The relationship still
provides meaningful prior evidence.

```
Example:
  Reported: "peripheral neuropathy"
  Existing: "diabetes mellitus"
  -> SYMPTOM_MODERATE
```

SYMPTOM_NONSPECIFIC
The relationship is real but occasional, indirect, or highly nonspecific,
so it does not provide meaningful prior evidence.

```
Example:
  Reported: "fatigue"
  Existing: "iron-deficiency anemia"
  -> SYMPTOM_NONSPECIFIC
```

DIFFERENT
No relationship stronger than sharing a broad organ, field, disease family,
laboratory context, risk factor, or similar wording.

```
Example:
  Reported: "platelet count"
  Existing: "eosinophil count"
  -> DIFFERENT
```

Use this decision order:

```
1. Same trait -> EXACT
2. One meaningfully contains the other -> PARENT or CHILD
3. Parallel traits sharing a direct biological or phenotypic axis -> CLOSE
4. One is a manifestation, consequence, complication, or marker of the other
   -> SYMPTOM
5. Otherwise -> DIFFERENT
```

The following relationships are strong enough to constitute meaningful prior
evidence:

```
EXACT
PARENT_FIRST
CHILD_FIRST
CLOSE_VERY
CLOSE_MODERATE
SYMPTOM_SPECIFIC
SYMPTOM_MODERATE
```

The following relationships are too broad, weak, or indirect to constitute
meaningful prior evidence:

```
PARENT_DISTANT
CHILD_DISTANT
CLOSE_VAGUE
SYMPTOM_NONSPECIFIC
DIFFERENT
```

Judge each entry independently. Do not rank the entries against each other.
Do not assume that an entry matches because it was retrieved by embedding
similarity.

Reply with one line per entry and nothing else:

```
<number> | <TAG>
```

Traits already linked to {gene}:

{traits}
