Heterozygosity is a measure of genetic diversity that describes how often different versions of a genetic sequence occur together or are encountered in a population. In a diploid organism, an individual is heterozygous at a locus when its two alleles differ. At the population level, heterozygosity may refer either to the observed proportion of heterozygous individuals or to the probability that two randomly sampled allele copies differ. These related measures are central to population genetics, evolutionary biology, conservation, and breeding.
Genetic basis and terminology
Diploid organisms generally carry two copies of each autosomal chromosome, one inherited from each parent. At a locus with alleles and , the genotypes and are homozygous, whereas is heterozygous. The alleles need not differ in their biological effects: a sequence difference can produce heterozygosity even when it has no detectable effect on the phenotype.
A locus may be a single nucleotide position, a repeated sequence, or a longer segment of DNA. Consequently, heterozygosity depends on how alleles and loci are defined. Two chromosome copies may be identical at one nucleotide but differ elsewhere in the same gene.
The distinction between heterozygosity and heterozygous state is useful. The latter describes a particular genotype; the former usually describes a frequency, probability, or genome-wide proportion.
Ordinary diploid genotype terminology does not apply unchanged to haploid regions, including some sex-chromosome regions. In organisms with polyploidy, more than two allele copies occur at a locus, and several definitions of heterozygosity are possible. For example, the proportion of individuals carrying more than one allele is not the same quantity as the probability that two allele copies differ.
Observed heterozygosity
Observed heterozygosity, commonly written , is the fraction of sampled diploid individuals that are heterozygous at a locus:
where is the number of heterozygous individuals and is the number successfully genotyped.
For example, if a sample of 100 individuals contains 30 , 50 , and 20 individuals, then
Across multiple loci, researchers may report the mean of locus-specific values. At the individual level, multilocus heterozygosity is often calculated as the number of heterozygous loci divided by the number of loci successfully assessed.
These summaries require an explicit denominator. Heterozygosity calculated only over known variable sites can be much higher than heterozygosity calculated over all reliably assessed genomic positions.
Expected heterozygosity
Expected heterozygosity, usually written , is calculated from allele frequencies:
where is the frequency of allele , and is the number of alleles at the locus.
This is the probability that two independently sampled allele copies differ. It is also commonly called gene diversity, although the locus need not be a protein-coding gene.
Under Hardy–Weinberg equilibrium, this probability equals the expected proportion of heterozygous individuals. For a locus with two alleles at frequencies and ,
In the preceding example, the allele frequencies are
Thus,
The observed value, , is close to the expectation, . Such agreement alone does not establish that all assumptions of Hardy–Weinberg equilibrium hold.
Limits and sample correction
For a biallelic locus, expected heterozygosity cannot exceed , reached when both alleles have frequency . For a locus with alleles, the maximum occurs when all are equally frequent:
Observed heterozygosity has different limits: every individual in a sample can be heterozygous, giving , even at a biallelic locus.
Calculating gene diversity directly from sample allele frequencies introduces a downward sampling bias. Under simple random sampling of independent allele copies, a common correction is
For completely genotyped diploid individuals, . Relatedness, population structure, and nonstandard sampling designs can require different treatment.
Evolutionary processes affecting heterozygosity
Mutation, drift, and population size
Mutation introduces new alleles, while genetic drift changes allele frequencies through random sampling between generations. In a finite population without mutation or immigration, drift ultimately removes variation at a neutral locus.
In an idealized diploid population, the expected decline in heterozygosity is approximately
where is initial heterozygosity, is the number of generations, and is effective population size. This expression assumes constant , neutrality, and no mutation or migration.
Effective population size describes the strength of genetic sampling processes and is not necessarily equal to the number of living individuals. Unequal reproductive success and fluctuations in population size can make the two substantially different.
Gene flow and selection
Gene flow can introduce alleles into a population and increase local heterozygosity when incoming alleles add variation. Its effect depends on the source and recipient populations’ allele frequencies.
Natural selection can either reduce or maintain heterozygosity. A selective sweep may reduce variation near a favored allele, whereas forms of balancing selection can maintain multiple alleles. Selection against heterozygotes can instead promote their deficiency.
Heterozygosity is therefore not a direct measure of selection, nor does a high value by itself establish greater adaptive capacity.
Inbreeding and population structure
Inbreeding increases the probability that an individual’s two allele copies descend from the same ancestral copy. Under a standard inbreeding model,
where is an inbreeding coefficient defined relative to a reference population.
A related population statistic is often expressed as
although practical estimators account for sampling and the structure of the data. Positive values indicate a heterozygote deficit relative to the chosen reference; negative values indicate an excess.
A deficit does not uniquely demonstrate mating between relatives. Pooling groups with different allele frequencies can produce a deficit even when mating is random within each group. This is the Wahlund effect. Genotyping errors and selection can also produce departures from expectation.
Measurement and genomic interpretation
Heterozygosity has been measured using protein variants, repeat-length markers, single-nucleotide polymorphisms, and whole-genome DNA sequencing. The expansion from biochemical markers to genomic data greatly increased the number of loci that could be examined, but also introduced new sources of technical bias.
Important measurement considerations include:
- Marker selection: panels chosen because their sites are already known to vary do not represent a random sample of the genome.
- Sequencing depth: insufficient coverage can cause one allele of a heterozygote to be missed.
- Mapping and calling errors: misaligned reads or duplicated sequences can create apparent heterozygosity.
- Missing data: uneven exclusion of difficult regions can alter comparisons.
- Ploidy: diploid assumptions are inappropriate for haploid regions or higher-ploidy genomes.
Genome-wide heterozygosity is commonly reported as heterozygous sites per callable nucleotide position. A callable position is one for which the available data meet specified reliability criteria.
Nucleotide diversity, usually denoted , is closely related but not identical in its usual application. It measures the average number of nucleotide differences per site between pairs of sampled sequences. At an individual nucleotide site, its population value has the same form as expected heterozygosity; averaged over a region, it describes sequence-level diversity.
An individual’s heterozygosity measures differences between its own chromosome copies. Population nucleotide diversity measures differences among sampled copies more generally. The two need not be equal, particularly in inbred or structured populations.
Applications and interpretive limits
In conservation genetics, heterozygosity is used alongside allele counts, relatedness, and demographic information to characterize genetic variation. In breeding, it helps describe crosses, breeding populations, and changes associated with mating systems. In evolutionary studies, variation in heterozygosity across genomes and populations contributes to analyses of population history, drift, and selection.
Several limitations constrain these applications:
- Rare alleles contribute relatively little. Because common allele frequencies dominate , the loss of rare alleles may substantially reduce allele richness while changing heterozygosity only modestly.
- Averages can conceal genomic structure. Individuals with similar overall heterozygosity can differ in the number and length of homozygous chromosome segments.
- Marker classes are not directly interchangeable. Highly variable repeat markers and biallelic nucleotide markers have different possible values and evolutionary behavior.
- Neutral diversity is not equivalent to functional diversity. Heterozygosity at sampled markers does not directly reveal variation affecting particular traits.
- Biological effects are context-dependent. Heterozygosity is not synonymous with heterozygote advantage or hybrid vigor, and a genome-wide value is not a direct measure of health or reproductive success.
Reported values are most interpretable when accompanied by the sampling design, marker definition, genomic denominator, population reference, and method used to estimate uncertainty.