Source: https://guthlabs.ai/docs/instruments/punnett-derived-carrier-risk-calculator-education/mechanism/

instrument mechanism / verified-clean

# Punnett-derived Carrier Risk Calculator (education) mechanism

The inspectable method record for Punnett-derived Carrier Risk Calculator (education), kept separate from implementation and qualification.

Status: **VERIFIED_CLEAN**Lifecycle: **verified-clean**

## Model scope

The existing educational and not-medical-advice label is appropriate, but the product must explicitly define genotype encoding, the dominant-trait heterozygote/carrier terminology, accepted input forms, and the exact output formulas before implementation can be verified as PASS.

- The specification does not provide explicit genotype-probability formulas or the genotype encoding needed to verify implementation.

- For autosomal recessive inheritance, the mapping of AA, Aa, and aa to unaffected, carrier, and affected must be explicitly defined.

- For autosomal dominant inheritance, “carrier” is ambiguous and is not generally a clinically appropriate category; the specification must state whether carrier_probability means heterozygous_probability (Aa).

- Parental “carrier status” alone is insufficient for some calculations unless the full parental genotype or the remaining genotype probabilities are supplied.

- The model needs an explicit assumption that each parent contributes one allele independently and that the trait follows complete, simple Mendelian inheritance without penetrance variation, de novo variants, allele-frequency inference, or linkage effects.

## Units and conventions

- Probabilities are dimensionless fractions from 0 to 1; multiply by 100 for percentages.

- Genotypes use two allele symbols, with the first allele conventionally written uppercase when representing the trait allele: autosomal recessive affected=aa, carrier=Aa, unaffected=AA.

- For autosomal dominant examples, A is the affected allele and a is the unaffected allele: affected=AA or Aa, unaffected=aa.

- For dominant examples only, heterozygous_probability is reported as the probability of Aa; this must not be labeled clinically as carrier probability without an explicit product decision.

- Inputs are complete parental genotypes, not merely labels such as “carrier” or “affected.”

- No physical units or unit conversions apply to Mendelian probabilities.

## Inspectable vector

One of 4 source-bound review vectors:

````
{
  "inputs": {
    "inheritance_pattern": "autosomal_recessive",
    "parent1_genotype": "AA",
    "parent2_genotype": "aa"
  },
  "expected_outputs": {
    "affected_probability": 0,
    "carrier_probability": 1,
    "unaffected_probability": 0
  },
  "working": "Parent 1 can contribute only A: P(A)=1. Parent 2 can contribute only a: P(a)=1. Offspring genotype: Aa with probability 1×1=1. Therefore affected aa = 0, carrier Aa = 1, unaffected AA = 0."
}
````

## Failure boundary

**BREAK_CLAIMED** · SELF_FLAGGED

The specification does not define behavior for invalid genotypes, does not uniquely define the unaffected category in recessive outputs, and permits dominant affected-status inputs that conceal an AA-versus-Aa distinction. These cases can produce rejection, undefined results, or materially different probabilities.

## Other records

[01**Overview**](<https://guthlabs.ai/docs/instruments/punnett-derived-carrier-risk-calculator-education/>)[03**Implementation**](<https://guthlabs.ai/docs/instruments/punnett-derived-carrier-risk-calculator-education/implementation/>)[04**Qualification**](<https://guthlabs.ai/docs/instruments/punnett-derived-carrier-risk-calculator-education/qualification/>)

## Sources and evidence

These pointers bind the page to evidence. They do not grant authority or collapse verification into deployment.

- `VERIFIED_LEDGER.json`

- `evidence-tranche5/tranche5-results.json`

- `BOUNDARY_HAZARD_CATALOG.json`