# Reading Pedigrees: Human Inheritance

> Fitting predictions support a model; one contradiction can rule it out. Under the simple models used here, you will read the symbols, test five modes of inheritance and name the observation that excludes each one. Then you will calculate the 2/3 carrier risk of an unaffected sibling, update a carrier probability from unaffected sons, and explain why penetrance, new mutations, mosaicism, heteroplasmy and small families limit certainty.

Course: Genetics and Molecular Biology (Inheritance). Level: core university (third year). About 20 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/pedigrees/

Builds on: Mendel's Peas: Segregation and Independent Assortment · Probability, Crosses and the Chi-Square Test · Extending Mendel: Dominance, Multiple Alleles and Gene Interaction

## What you'll learn

- Read pedigree symbols and recognize the signatures of autosomal dominant, autosomal recessive, X-linked recessive, X-linked dominant and mitochondrial inheritance, naming the observation that rules each one out.
- Calculate carrier and recurrence probabilities under stated assumptions, including the 2/3 carrier probability of an unaffected sibling of an affected person.
- Update a carrier probability with Bayes' rule from the phenotypes of a woman's sons.
- Explain how reduced or age-dependent penetrance, new mutations, germline mosaicism, heteroplasmy and small families make pedigree inference uncertain.

Human geneticists cannot set up crosses, so they work backwards from existing families. A **[pedigree](https://stemhop.com/methods/pedigree-analysis/)** is a family tree drawn to a shared code.

> **The one idea.** Fitting predictions support a model; one contradiction can rule it out.
>
> *Scope:* the five simple, single-gene models used here, assuming full **penetrance** (everyone with the genotype shows the trait), no new mutation and correct family relationships. A model that fits is supported, not proved.

## Read the evidence first

The first film used three symbols; the full code has more. In the 2022 revision of the National Society of Genetic Counselors standard, a square is a man or boy, a circle is a woman or girl, and a diamond is a nonbinary person or someone whose sex or gender is not specified. The revision ties shape to gender and records sex assigned at birth separately. A filled symbol means affected by the condition named in the key; a diagonal slash marks someone who has died.

A horizontal line joins partners. A vertical line drops to a horizontal sibship line, from which each child hangs. A double partner line marks **consanguinity**: the partners are related by descent, as first cousins are. Roman numerals label generations, and people are numbered left to right, so II-3 identifies one person. An arrow marks the **proband**, through whom the family came to attention.

A **carrier** is heterozygous for a recessive allele and usually does not show the trait. Many textbook charts put a central dot in the circle of a woman known to carry an X-linked recessive allele. The 2022 revision shows carrier status differently; check the key.

**Predict.** Two partners are joined by a double line. Can you conclude that both carry the allele being studied?

> **Check your prediction.** No. The line records shared ancestry, not genotype. Consanguinity raises the chance that both carry the same rare allele but does not establish carrier status or prove recessive inheritance.

## Test five inheritance models

| Model | Prediction under the simple model | A contradiction |
| --- | --- | --- |
| **Autosomal dominant** | One copy in a heterozygote is enough. The trait usually appears in successive generations and both sexes; father-to-son transmission is possible. An affected heterozygote and unaffected partner give each child a 1/2 allele risk. | An affected child has two unaffected parents. |
| **Autosomal recessive** | Affected people have two pathogenic alleles at one locus, identical or different. Unaffected carrier parents can have affected children of either sex, so cases often cluster among siblings, with few affected relatives elsewhere. Shared ancestry is more common because relatives are likelier to carry the same rare allele. | Two affected parents have an unaffected child, if the same gene causes both parents’ trait. |
| **X-linked recessive** | A male has one X, so one copy can affect him. Affected men are much more common than affected women. An affected man passes the allele through carrier daughters to half their sons. Heterozygous women can show signs, partly through X-inactivation. | Demonstrated father-to-son transmission. An affected father-son pair establishes this only if the mother carries no copy. |
| **X-linked dominant** | An affected father passes the allele to every daughter and no son. An affected heterozygous mother passes it to half her children of either sex. | An affected father has an unaffected daughter or transmits the allele to a son. |
| **Mitochondrial** | Human mitochondrial DNA normally comes from the mother. An affected mother can have affected children of both sexes; an affected father transmits the trait to none. | Transmission from an affected father. |

Recessive deafness shows why that row’s qualification matters: different genes can cause deafness, so two deaf parents can have a hearing child. Mitochondrial DNA encodes only a handful of respiratory-chain subunits (see [oxidative phosphorylation](https://stemhop.com/lessons/oxidative-phosphorylation/)); most mitochondrial proteins are nuclear-encoded, so many mitochondrial diseases instead follow Mendelian patterns.

*Figure (AI-generated illustration).* A contradiction rules out a simple model under its assumptions: an affected child of two unaffected parents rules out autosomal dominant; an unaffected child of two affected parents rules out autosomal recessive when the same gene causes both parents' trait; father-to-son transmission, with a mother who carries no copy, rules out X-linked recessive. Image: three boxed pedigree panels labeled Autosomal dominant, Autosomal recessive, and X-linked recessive. Left: two unaffected parents with an affected daughter, captioned affected child, unaffected parents. Middle: two affected parents with an unaffected son, captioned unaffected child, affected parents. Right: an affected father and unaffected mother with an affected son, the line from father to son labeled Father to son.

**Predict.** An affected father has an unaffected daughter. Under the scope above, which of the five models does that one observation rule out?

> **Check your prediction.** Only X-linked dominant, because every daughter receives her father’s X. The other four models all allow an unaffected daughter.

Do not use “skips a generation” as proof of recessive inheritance; reduced penetrance can make a dominant allele appear to skip. Rett syndrome and fragile X syndrome, introduced in [synapse disorders](https://stemhop.com/lessons/synapse-disorders/), are X-linked examples complicated by new mutations and repeat expansion.

Once a model survives those tests, the next question is how strongly each possible genotype is supported.

**Film transcript: Evidence Reweights the Possibilities**

### Film transcript: Evidence Reweights the Possibilities

The first film tested inheritance models against family observations. How can learning that someone is unaffected raise their carrier probability? The surprise comes from comparing only genotypes still compatible with the observation.

Begin with every genotype the inheritance model allows. Give each possibility its starting weight before using the person's phenotype. These weights come from transmission rules and any population assumptions. Write them down before looking at the answer you hope to get. Now condition on what you observed. Any genotype incompatible with that observation loses its weight. Compatible possibilities remain. This is a filter, not a change to the person's cells. Keep the survivors in their original proportions. Suppose a new observation excludes one visible possibility. Which tiles should remain in the probability bar, and what must happen next? Watch the total bar, not just the disappearing tile.

Remove the incompatible tile. Then stretch the survivors to fill the whole bar again. That second step is renormalizing. The surviving possibilities now divide all the probability between them. A carrier tile can therefore occupy more of the remaining bar. This can happen even when the person is unaffected. The comparison set became smaller. Nothing happened to the person's DNA.

Evidence changes genotype probabilities, not genotypes. Only the description of your uncertainty changed. Sometimes evidence does not eliminate a hypothesis. It is merely more likely under one hypothesis than another. Bayes' rule handles that case. The observation changes weights without making either hypothesis impossible. Likelihood measures that predictive fit. Each hypothesis starts with a prior weight. Multiply that weight by how well the hypothesis predicts the observation. A better prediction earns a larger updated weight. A weak prediction earns less. Finally, renormalize those weighted hypotheses. Their new shares form the posterior: the probabilities after accounting for the evidence. Together, those shares still fill the complete probability bar.

More observations repeat the same cycle. Start with current weights, ask what each hypothesis predicts, apply the evidence, then renormalize. Each new fact becomes the next filter. The method remains the same. The answer still depends on the model. Penetrance, new mutation, family relationships, and population assumptions all shape those starting weights and likelihoods. If an assumption changes, recompute instead of reusing the old answer. With that picture, the text can calculate exact carrier and recurrence probabilities. Keep the filter, the weights, and the assumptions visible. The worked examples begin with different family evidence.

From the film

- Evidence changes genotype probabilities, not genotypes.
- Conditioning removes incompatible outcomes and renormalizes the outcomes that remain.
- Bayes' rule weights each hypothesis by how well it predicts the evidence.
- Risk calculations are only as sound as their stated inheritance assumptions.

## Conditional probability: the 2/3 rule

The second film reweighted schematic tiles; now put exact numbers on them. A woman’s brother has cystic fibrosis, an autosomal recessive disorder caused by two pathogenic *cystic fibrosis transmembrane conductance regulator (CFTR)* alleles. Her parents and she are unaffected. Her partner has no family history. What is the probability that their first child is affected?

Within the scope above, assume a partner carrier frequency of 1/25, a figure often quoted for northern European ancestry and lower in many populations; it is an exercise assumption, not a universal number.

Her affected brother shows that both parents are carriers, Aa. Before her phenotype is used, Aa x Aa gives four equally likely outcomes: AA, Aa (A from the mother), aA (A from the father) and aa, each 1/4.

**Predict.** She is unaffected. Which outcome disappears, and what fraction of the remaining outcomes are carriers?

> **Check your prediction.** The affected aa outcome disappears. Of the three equal outcomes left, Aa and aA are carriers:
>
> > P(carrier | unaffected) = (1/2) / (1/4 + 1/2) = (1/2) / (3/4) = 2/3

If both partners are carriers (her partner with probability 1/25), both pass the recessive allele with probability 1/4. Therefore

> P(affected child) = 2/3 x 1/25 x 1/4 = 2/300 = 1/150

As a decimal check, 1/150 is about 0.0067, and 0.667 x 0.04 x 0.25 = 0.0067. The tempting 1/2 was her carrier probability before her unaffected phenotype removed one outcome. These are logic exercises, not risk figures for a real family.

*Interactive on the lesson page:* Pedigree risk calculator: carrier and recurrence odds

## Updating with evidence: Bayes and unaffected sons

Here evidence reweights the hypotheses instead of deleting one, so Bayes’ rule from the second film applies. A woman’s mother is a known carrier of a fully penetrant X-linked recessive allele, and her father is unaffected. The untested woman has two unaffected sons.

Her unaffected father gave her a normal X and her mother one of two X chromosomes, so her prior is 1/2 carrier and 1/2 not carrier.

**Predict.** Are two unaffected sons more likely if she is a carrier or if she is not? Should her carrier probability rise or fall?

> **Check your prediction.** They are more likely if she is not a carrier, so her carrier probability falls. If she is a carrier, each son is unaffected with probability 1/2, so the likelihood for two is 1/4. If she is not a carrier, the likelihood is 1.

Multiply each **prior** (the weight before the evidence) by its **likelihood** (the chance of the evidence under that hypothesis): carrier gives 1/2 x 1/4 = 1/8; not carrier gives 1/2 x 1 = 1/2. Normalize the two weights:

> P(carrier | two unaffected sons) = (1/8) / (1/8 + 1/2) = 1/5

Prior odds of 1 : 1 become posterior odds of 1 : 4, because the evidence is four times likelier if she is not a carrier. Under the same model, her next son’s affected risk is 1/5 x 1/2 = 1/10.

## Why real pedigrees mislead

One contradiction rules a model out only inside the scope, and real families break the scope in four ways.

- **Genotype may not predict one visible phenotype.** Huntington disease is dominant, but symptoms usually begin in adulthood, so a young unaffected child of an affected parent may carry the allele. Penetrance also depends on CAG repeat length in *HTT*: about 36 to 39 repeats is usually described as reduced penetrance, and 40 or more as full penetrance. **Variable expressivity** means the same genotype produces different clinical pictures.
- **A variant may be new or mosaic.** A new pathogenic variant can produce a dominant disorder in a child of unaffected parents, which looks exactly like the observation that rules out autosomal dominant inheritance. In **germline mosaicism**, a parent carries a variant in only some cells, including some sperm or eggs. In a study of 58 families with one affected child and an apparently new variant, deep sequencing of several parental tissues estimated recurrence below 0.1% in 35 families; 5 had parental mosaicism associated with higher risk, and three fathers had the variant in 5.6 to 12.1% of sperm DNA.
- **Mitochondrial genotype can be a mixture.** With **heteroplasmy**, a cell contains normal and mutant mitochondrial DNA. A mother can transmit different proportions to different children, so severity varies and some children may be unaffected.
- **A family may be too small to reveal the pattern.**

**Predict.** If two carrier parents have only two children, is seeing no affected child surprising enough to reject autosomal recessive inheritance?

> **Check your prediction.** No. Each child has a 3/4 chance to be unaffected, so the chance that both are unaffected is (3/4)2 = 9/16, more than half the time. Small families can look as though nothing is inherited.

Apparent exceptions also need molecular checks: among 11,035 trios, all seven apparent paternal mitochondrial signals were explained by a mitochondrial-derived nuclear insert shared by father and child.

### Design problem

An affected woman has five children, all affected. Her affected son has three children, none affected. Her affected daughter has two children, both affected. Severity varies. Decide between mitochondrial and X-linked dominant inheritance, and identify data that would settle it.

> **One way to do it.** **Test the affected son.** He transmits no mitochondrial DNA, so three unaffected children fit mitochondrial inheritance. Under fully penetrant X-linked dominant inheritance every daughter receives his affected X, so they exclude that model only if at least one is a daughter.
>
> **The rest does not discriminate.** Both models fit an affected mother with affected children. A mother transmits mitochondria to every child; under X-linked dominant inheritance, five affected children of a heterozygous mother has probability (1/2)5 = 1/32: possible but unlikely.
>
> **Variable severity fits both models, so it cannot decide between them.** Under heteroplasmy children receive different mutant fractions; X-inactivation in women can vary an X-linked dominant phenotype.
>
> **Settle it molecularly.** An affected father with an affected daughter whose mother lies outside the maternal line and lacks the family’s mitochondrial variant would exclude mitochondrial inheritance. Or sequence affected and unaffected relatives, seek a mitochondrial variant confined to the maternal line, and measure heteroplasmy in several tissues. Heteroplasmy tracking severity, with the variant absent from the son’s children, would support the mitochondrial model.

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/pedigrees/
