Principles of Inheritance: Mendelian and Non-Mendelian Patterns

Introduction

Genetics questions in finals are rarely about naming a gene. They are about recognising a pattern of inheritance from a pedigree or a clinical vignette, and using it to predict recurrence risk. That skill rests on a small number of principles, most of them dating back to Mendel, plus a handful of well-defined exceptions that examiners are fond of because they catch out anyone who has only learned the simple rules.¹

This article sets out the vocabulary and the classic Mendelian patterns, then works through the situations where the simple rules break down: incomplete penetrance, variable expressivity, anticipation, imprinting, mosaicism and mitochondrial inheritance. The specific syndromes referenced as examples - Down syndrome, Marfan syndrome and so on - each have their own article; this one is about the logic that ties them together.

A working vocabulary first. A gene is a sequence of DNA encoding a product; its alleles are the alternative versions found at that position (locus) on homologous chromosomes. The genotype is the pair of alleles an individual carries at a locus; the phenotype is the observable result. Homozygous means both alleles are identical, heterozygous means they differ, and hemizygous describes a locus present in only one copy, as with X-linked genes in a male.

Autosomal dominant inheritance

A single copy of the variant allele on one of the 22 autosomes is enough to produce the phenotype. The classic pedigree hallmark is that affected individuals appear in every generation, and an affected person has, on average, a 50% chance of passing the condition to each child, regardless of the child's sex.

• Male-to-male transmission occurs, which is the key feature distinguishing autosomal dominant from X-linked dominant disease
• Unaffected individuals do not transmit the condition, in the absence of reduced penetrance
• Two affected heterozygous parents have a 1 in 4 chance of a homozygous child, which is often more severely affected than either parent
• Many autosomal dominant conditions arise from loss of one functional copy of a gene whose product is needed in a critical amount - haploinsufficiency - or from a variant protein that actively interferes with the normal one, a dominant-negative effect

Examples include Marfan syndrome, neurofibromatosis type 1, Huntington's disease, familial hypercholesterolaemia, and adult polycystic kidney disease. Several of these are covered in their own articles, since the pattern of inheritance is only the starting point for the clinical picture.

Autosomal recessive inheritance

Both copies of the gene must carry a variant for the phenotype to appear. Heterozygous carriers are clinically unaffected because the single normal allele produces enough functional product. Most autosomal recessive conditions are caused by loss-of-function variants in enzymes, where roughly half the normal enzyme activity is still sufficient for normal function.

• Typically affects siblings, not parents - the pedigree looks as if the condition has skipped a generation
• Two carrier parents have a 1 in 4 (25%) risk of an affected child, a 1 in 2 (50%) chance of a carrier child, and a 1 in 4 chance of a child with neither variant
• Consanguinity substantially increases risk, because related parents are more likely to share the same rare recessive allele inherited from a common ancestor
• Equal sex distribution, since the gene is autosomal

Examples include cystic fibrosis, sickle cell disease, phenylketonuria, most inborn errors of metabolism, and haemochromatosis (which is unusual among recessive conditions in presenting in mid-adult life rather than childhood).

X-linked inheritance

Genes on the X chromosome show a distinctive pattern because males are hemizygous - they have only one X, inherited from their mother, and one Y, inherited from their father.

X-linked recessive

• Affects males far more often than females, since a male expresses the phenotype with a single variant allele while a female needs both copies affected
• No male-to-male transmission - a father passes his Y chromosome, not his X, to his sons
• All daughters of an affected male are obligate carriers
• A carrier mother has a 50% chance of passing the variant to each child; affected sons and carrier daughters occur with equal probability among her children
• Skips generations, transmitted through unaffected or mildly affected carrier females

Examples include Duchenne muscular dystrophy, haemophilia A and B, and G6PD deficiency. Carrier females can show mild features through skewed X-inactivation, where the X carrying the normal allele is preferentially inactivated in a higher proportion of cells than the usual random 50%.

X-linked dominant

Rarer than X-linked recessive. A single variant allele causes disease in either sex, but affected fathers pass the condition to all their daughters and none of their sons, which is the pattern that distinguishes it from autosomal dominant inheritance. Some X-linked dominant conditions, such as Rett syndrome and incontinentia pigmenti, are lethal in hemizygous males, so affected individuals are almost exclusively female, with the pregnancy loss of affected male fetuses inflating the apparent female preponderance.

Y-linked inheritance

Very rare in clinical practice, largely limited to some causes of male infertility. Transmission is exclusively father to son, since only males carry a Y chromosome, giving a pedigree with affected males in every generation and no female involvement at all.

Reading a pedigree

Pedigree questions reward a systematic approach rather than pattern-spotting from memory.

1. Count generations and roughly what fraction of each generation is affected. A high proportion affected, in every generation, suggests dominant inheritance; a low proportion, clustered in siblings, suggests recessive.
2. Look specifically for male-to-male transmission. If present, the condition cannot be X-linked, since a father cannot pass an X-linked allele to a son.
3. Compare the sexes affected. A marked excess of affected males points to X-linked recessive; roughly equal numbers point to autosomal inheritance.
4. Check whether unaffected individuals ever transmit the condition. This is expected in recessive and X-linked recessive disease (via carriers) but should not happen in straightforward autosomal dominant disease, unless non-penetrance is at play.
5. Note consanguinity, usually shown by a double line joining parents, which raises the prior probability of recessive inheritance.
6. Only once the pattern is established, calculate the numerical risk using the relevant Mendelian ratio for the specific relationship asked about.

Penetrance and expressivity

Two related but distinct concepts frequently confused with each other, and a favourite short-answer distinction.

• Penetrance is the proportion of people with a given genotype who show any phenotype at all. It is an all-or-nothing, population-level statistic: BRCA1 pathogenic variants have around 60-70% penetrance for breast cancer by age 70, meaning roughly a third of carriers never develop the disease.
• Expressivity describes the severity or range of features in someone who does express the phenotype. Neurofibromatosis type 1 is highly penetrant - almost everyone with the variant shows some feature - but hugely variable in expressivity, from a handful of cafe-au-lait patches to disfiguring plexiform neurofibromas.
• Reduced (incomplete) penetrance explains an apparently skipped generation in an otherwise dominant pedigree: the intervening individual carries the allele but never developed the phenotype, and can still transmit it to their own children.

Anticipation and trinucleotide repeat disorders

Anticipation is the tendency for a genetic disease to present earlier and more severely in successive generations. It is characteristic of disorders caused by expansion of an unstable trinucleotide (or larger) repeat sequence, where the repeat tends to lengthen further with each transmission, particularly through spermatogenesis in some conditions.

Genomic imprinting

Most autosomal genes are expressed from both the maternal and paternal copy. A small number of genes are imprinted: one parental copy is epigenetically silenced (by DNA methylation, established in the germline), so only the other parent's allele is normally expressed. Loss of the active copy therefore causes disease, even though the silenced copy is structurally normal.

The clearest illustration is chromosome 15q11-13, which contains reciprocally imprinted genes. The same microdeletion produces a completely different syndrome depending on which parent it is inherited from:

• Deletion of the paternal copy - the paternally expressed gene is lost, causing Prader-Willi syndrome: neonatal hypotonia and poor feeding, followed by hyperphagia and childhood obesity, learning disability and hypogonadism
• Deletion of the maternal copy - the maternally expressed UBE3A is lost, causing Angelman syndrome: severe developmental delay, absent speech, ataxia, seizures and a characteristic happy, easily excitable demeanour

Mosaicism

Mosaicism is the presence of two or more genetically distinct cell lines in one individual, arising from a mutation occurring after fertilisation, so that only a proportion of cells carry it. The clinical severity depends on which tissues are affected and what fraction of cells carry the variant.

• Somatic mosaicism affects body tissues and can produce patchy or milder disease than the fully penetrant form - for example some cases of neurofibromatosis type 1 or McCune-Albright syndrome
• Gonadal (germline) mosaicism is confined to the gametes. Parents are phenotypically normal and test negative for the variant on a blood sample, yet can have more than one affected child, because the mutation is present in a subset of egg or sperm cells but absent from the somatic cells that were tested
• Gonadal mosaicism is the usual explanation when a supposedly de novo autosomal dominant condition, such as osteogenesis imperfecta, recurs in a second sibling of unaffected, genetically-tested parents

Mitochondrial inheritance

Mitochondria carry their own small circular genome, present in many copies per cell, encoding proteins of the oxidative phosphorylation chain. Mitochondria are inherited exclusively from the mother: the ovum contributes essentially all of the mitochondria to the zygote, while sperm mitochondria are destroyed after fertilisation.

• An affected mother transmits the condition to all of her children, of either sex
• An affected father transmits it to none of his children
• Heteroplasmy - a cell contains a mixture of normal and mutant mitochondrial DNA in varying proportion - explains why severity and organ involvement vary widely between individuals, and even between siblings, depending on the proportion of mutant mitochondria that happened to segregate into their tissues during development
• Tissues with the highest energy demand - the central nervous system, cardiac and skeletal muscle - are typically affected earliest and most severely, since they are least tolerant of a shortfall in oxidative phosphorylation

Examples include Leber hereditary optic neuropathy and MELAS (mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes). These are rare but a reliable source of an SAQ, because the maternal-only transmission pattern is unambiguous once recognised.⁴

Multifactorial and polygenic inheritance

Most common disease does not follow a single-gene pattern at all. Multifactorial (complex) inheritance describes traits and diseases resulting from the combined, additive effect of many genetic variants, each of small individual effect, interacting with environmental factors. Type 2 diabetes, hypertension, most congenital heart disease, neural tube defects and schizophrenia all fall into this category.

• Recurrence risk cannot be calculated from Mendelian ratios; it is estimated empirically, from observed recurrence rates in family studies
• Risk to relatives falls off sharply with genetic distance - much higher in first-degree relatives than second-degree, unlike single-gene disease where the risk to any given relative depends only on their relationship, not the population baseline
• Risk rises with the number of affected relatives and with the severity of disease in the index case
• Some conditions show a sex-dependent threshold effect: pyloric stenosis is commoner in males, but when it does occur in a female (who needed a higher liability of predisposing factors to cross the threshold), her relatives carry a higher recurrence risk than the relatives of an affected male

Chromosomal versus single-gene disease

It is worth being explicit about a distinction implied throughout this article. Everything above describes single-gene (Mendelian) disease, where the abnormality lies within one gene. Chromosomal disorders instead involve gain, loss or rearrangement of whole chromosomes or large segments of them - trisomies such as Down syndrome, sex chromosome aneuploidies such as Turner and Klinefelter syndrome, and microdeletion syndromes. These generally arise from errors in meiosis (most often non-disjunction) rather than being transmitted according to Mendel's laws, and each has its own recurrence risk that depends on the mechanism involved, covered in the individual articles on those conditions.

Clinical application

In practice, establishing the inheritance pattern feeds directly into genetic counselling: quoting an accurate recurrence risk, deciding who else in the family should be offered testing or surveillance, and informing reproductive decisions such as prenatal diagnosis or pre-implantation genetic testing. The mechanics of that conversation, and the tests used to confirm a diagnosis, are covered in Genetic Counselling and Testing.

• Always draw or interpret a three-generation pedigree before quoting a risk figure³
• State risks as they would be understood by a patient - '1 in 4', not just '25%' or '0.25' - and check understanding
• Distinguish the risk of being a carrier from the risk of being affected, which is a common point of confusion for patients and for exam answers alike
• Remember that a normal test result changes risk but does not always reduce it to zero, particularly where non-penetrance, mosaicism or a as-yet-unidentified causative variant is possible