Carrier probability calculator: Bayesian risk assessment from pedigree and population

Short version. Carrier probability is the probability that a clinically unaffected individual carries one copy of a disease allele and can transmit it to offspring. A calculator combines three ingredients: a prior (from pedigree, from population carrier frequency, or both), a likelihood (the probability of observed clinical and test evidence given carrier vs non-carrier), and normalisation. The output is a posterior carrier probability that updates as more evidence becomes available. Correct use of population carrier frequencies requires ancestry-appropriate data. Consanguinity increases carrier-overlap probability between partners. Negative molecular testing reduces but does not usually eliminate residual risk, because no panel has 100% sensitivity.

What carrier probability actually represents

Carrier probability is a conditional statement: given the information we have, what is the probability that this specific individual is heterozygous (or hemizygous for X-linked conditions) for the disease allele? For autosomal recessive and X-linked recessive conditions, the carrier is almost always clinically unaffected. The question matters because carriers can transmit to offspring, and the reproductive risk for a couple depends on both partners' carrier probabilities.

"Carrier" as a word is also used in hereditary cancer contexts (a BRCA1 "carrier"), where the inheritance is autosomal dominant and the carrier is at risk of manifesting cancer themselves. That is a different usage; the Bayesian framework is similar but the interpretation of the probability is different. This page focuses on recessive-style carrier probability, with a brief note on the dominant case below.

Priors from the pedigree

For a relative of an affected proband, the prior carrier probability is determined by Mendelian transmission along the pedigree.

For X-linked recessive, the pedigree-based priors for female relatives mirror the X-transmission rules: 1/2 for the daughter of an obligate carrier mother, 1 for the daughter of an affected father, and so on.

Priors from population carrier frequency

When a partner has no family history, their prior carrier probability comes from population data. The Hardy-Weinberg equilibrium gives a useful link for rare alleles: if the disease incidence is q², then the carrier frequency 2pq ≈ 2q. So for cystic fibrosis at incidence roughly 1 in 2,500 in Northern Europeans, q ≈ 1/50 and carrier frequency ≈ 1/25.

Several conditions have population-stratified carrier frequencies. Examples include:

• Cystic fibrosis — elevated in Northern European populations.
• Tay-Sachs, Canavan, Gaucher type 1, familial dysautonomia, and several others — elevated in Ashkenazi Jewish populations, historically the subject of dedicated carrier screening panels.
• Beta-thalassaemia — elevated in Mediterranean, Middle Eastern, and South Asian populations.
• Sickle cell disease — elevated in sub-Saharan African, Caribbean, and some Mediterranean and Middle Eastern populations.
• Spinal muscular atrophy — pan-ethnic with modest population variation.

Using the right population data matters. A carrier frequency of 1 in 25 vs 1 in 100 translates to a fourfold difference in couple risk. Honest reporting uses ancestry-appropriate frequencies where available and describes uncertainty qualitatively where exact figures are disputed. Ancestry is also sometimes mixed, in which case a weighted combination of frequencies can be used.

Bayesian updating with observed evidence

The Bayesian framework is simple in principle:

Posterior ∝ Prior × Likelihood

The likelihood is the probability of the observed evidence given each hypothesis. Examples:

• Unaffected offspring: for an at-risk carrier mother in an X-linked pedigree, each unaffected son gives likelihood 1/2 under the carrier hypothesis and approximately 1 under the non-carrier hypothesis.
• Unaffected siblings of an AR proband: no update to the 2/3 prior unless the sibling has had testing themselves; their unaffected phenotype is already part of the prior.
• Molecular test result: a targeted variant test for a known familial variant is essentially deterministic (assuming reliable technique); panel screening has non-trivial false-negative rates depending on panel coverage.
• Biochemical evidence: factor VIII level in a haemophilia A carrier is informative but not deterministic; SMN1 copy-number testing is deterministic for most SMA carriers but 2+0 silent carriers complicate the picture.

A well-designed calculator presents the intermediate probabilities transparently so the counsellor can see which evidence drove the final number.

Consanguinity and couple risk

Consanguinity increases the joint probability that both partners carry the same rare allele inherited from a common ancestor. The couple risk for a rare recessive condition is approximately Fq + (1 - F)q², where F is the inbreeding coefficient and q is the allele frequency. For a first-cousin couple (F ≈ 1/16) considering a very rare condition (q = 1/1000), Fq dominates at approximately 1/16000, far above the q² = 1/1000000 population-baseline risk.

When one partner has a known family history and the other is a relative, the pedigree-based prior already captures the consanguinity effect for that specific allele. The population-baseline contribution should not be double-counted. Evagene's consanguinity-aware calculation handles this automatically.

The BRCAPRO case: dominant carrier probability

For hereditary cancer syndromes such as HBOC (BRCA1/BRCA2), "carrier" means carrier of a pathogenic variant that itself confers elevated cancer risk in a dominant fashion. The Bayesian framework remains, but the likelihood model is empirical: BRCAPRO uses published population allele frequencies and age-specific cancer-penetrance curves to compute a posterior BRCA1 and BRCA2 carrier probability given the observed family pattern. MMRpro and PancPRO do the same for Lynch syndrome and pancreatic cancer susceptibility, respectively. See our hereditary cancer risk assessment page for the detail.

The conceptual point: carrier probability is the right abstraction for both recessive and dominant inheritance, but the clinical meaning and the penetrance model differ.

How Evagene supports this

Evagene's Mendelian calculator computes carrier probabilities per-individual on the pedigree for autosomal recessive and X-linked recessive conditions annotated from the 200+ disease catalogue. Priors are derived from pedigree transmission; when population data are needed for a partner without family history, ancestry-appropriate carrier frequencies are applied where available and estimated from Hardy-Weinberg where necessary. Ancestry can be entered manually or inferred from 23andMe SNP imports when those data are present.

Bayesian updates incorporate observed unaffected or affected offspring, sibling status, and recorded molecular test results. For hereditary cancer risk specifically, BRCAPRO, MMRpro, and PancPRO run on the pedigree to give empirical carrier probabilities alongside the Mendelian reasoning. Consanguinity detection and Wright's coefficient calculation are automatic and are integrated into couple-risk estimates. Batch screening applies this analysis across the full disease catalogue for a proband, so a clinician can see at a glance which conditions have elevated carrier risk in this family.

The calculator is transparent about which pedigree features and which population frequencies drove a number — essential for a counsellor needing to justify the result in a clinic letter.

Frequently asked questions

Related reading

• Mendelian inheritance calculator overview
• Autosomal recessive calculator
• X-linked recessive calculator
• Consanguinity calculator
• Identifying inheritance patterns
• Hereditary cancer risk assessment
• Clinical genetics pedigree tools