Allele and Genotype Frequencies

A population's gene pool is the sum total of all alleles at all loci carried by every reproducing individual in that population. For a single locus with two alleles, call them A and a, every diploid individual carries one of three possible genotypes: AA, Aa, or aa. The genotype frequency of, say, AA is simply the number of AA individuals divided by the total number of individuals sampled — it answers "what fraction of organisms look like this genetically?" The allele frequency, by contrast, answers a different question: "what fraction of all the gene copies floating in this population's pool are the A allele?" Because each diploid individual carries two alleles at the locus, a population of N individuals contains 2N total allele copies. An AA homozygote contributes two A alleles to that pool; an aa homozygote contributes two a alleles; a heterozygote Aa contributes one of each. Allele frequency is conventionally denoted with lowercase letters — most commonly p for the frequency of the more common or reference allele, and q for the frequency of the alternative allele — and by definition p + q = 1 for a two-allele locus, since every gene copy must be one or the other. It is essential to keep genotype frequency and allele frequency conceptually separate even though they are computed from the same raw data. Genotype frequencies describe individuals; allele frequencies describe gene copies. A locus can have three genotype frequencies (for AA, Aa, aa) that must sum to 1, and simultaneously two allele frequencies (for A and a) that must independently sum to 1. These two sets of numbers are linked by simple counting arithmetic, not by any assumption about mating patterns, selection, or evolutionary equilibrium — that linkage becomes important only when we ask whether the observed genotype frequencies match a specific predicted pattern, which is the domain of the Hardy-Weinberg Principle.

Step 1: Sample the population and classify genotypes. Population geneticists cannot examine every individual in a species, so they take a representative sample of N diploid individuals and genotype each one at the locus of interest — historically by blood typing, protein electrophoresis, or Southern blotting; today overwhelmingly by direct DNA sequencing or genotyping arrays (SNP chips, PCR-based assays). Each individual is scored as AA, Aa, or aa, producing three counts that must sum to N.

Step 2: Compute genotype frequencies by simple division. If a sample of N = 1000 individuals contains

• 490 AA
• 420 Aa
• 90 aa

then the genotype frequencies are:

Note these three frequencies sum to 1.00, as they must — every individual has exactly one genotype.

Step 3: Count alleles, not individuals. This is the conceptual pivot point of the whole topic. Each AA individual carries two A alleles; each Aa individual carries one A and one a; each aa individual carries two a alleles. Total allele copies in the population = 2N = 2000. Total A alleles = (2 × number of AA) + (1 × number of Aa) = (2 × 490) + (1 × 420) = 980 + 420 = 1400 Total a alleles = (2 × number of aa) + (1 × number of Aa) = (2 × 90) + (1 × 420) = 180 + 420 = 600 Check: 1400 + 600 = 2000 = 2N. ✓

Step 4: Divide by total allele copies to get allele frequency.

Notice p + q = 1.00, confirming internal consistency. This same formula generalizes directly from genotype frequencies rather than counts, which is often more convenient:

Each heterozygote contributes "half an A" and "half an a" to the frequency calculation, because exactly half of its two alleles are A and half are a. This shortcut — homozygote frequency plus half the heterozygote frequency — is the single most useful formula in this entire topic and is worth memorizing.

Multi-allelic loci. Many real loci — the ABO blood group, HLA genes, microsatellite markers — have more than two alleles. The logic extends without modification: sum twice the frequency of each homozygous class plus the frequency of every heterozygous class that includes that allele, all divided by 2N. For three alleles A1, A2, A3 with six possible genotypes, the frequency of A1 is:

X-linked loci require an adjustment because males are hemizygous. In a population with equal numbers of males and females, a female contributes 2 allele copies but a male contributes only 1 (he has a single X). The denominator for total allele count becomes (2 × number of females) + (1 × number of males), not simply 2N, and this must be tracked separately for X-linked traits such as red-green color blindness or hemophilia A.

Allele and genotype frequency are, at first glance, "just arithmetic" rather than a hypothesis needing experimental proof — you cannot falsify a definition. But the historical case for treating the allele, rather than the whole organism or the whole genotype, as the fundamental unit that changes over generations was a genuine empirical and conceptual battle, and it is worth knowing because it explains why population genetics is built the way it is.

Reginald Punnett's 1908 puzzle and Hardy's and Weinberg's response. In the early 1900s, after the rediscovery of Mendel's work, many biologists — including the prominent Mendelian geneticist Reginald Punnett — assumed that a dominant trait must become more common over generations simply because it is dominant, since it is expressed whenever present. Punnett was publicly challenged by the mathematician G. H. Hardy's friend, statistician R. C. Punnett's own colleague, on why a dominant condition (brachydactyly, short fingers) had not swept to fixation in the British population despite being dominant for many generations. Punnett brought the question to Hardy, a Cambridge mathematician, and independently the German physician Wilhelm Weinberg addressed the identical problem in 1908. Both derived, using nothing more than basic algebra applied to random mating, that allele frequencies remain constant across generations in the absence of evolutionary forces, and that dominance has no intrinsic tendency to increase an allele's frequency. This was the crucial conceptual proof that frequency — not dominance, not visibility — is what evolution acts on and what must be tracked. Hardy and Weinberg's algebra depends entirely on being able to define p and q precisely, exactly as constructed above; the entire principle collapses without a rigorous allele-frequency bookkeeping system, which is why this arithmetic is taught as the prerequisite step.

Direct empirical confirmation: blood group surveys. The first large-scale test of whether real populations behave in a way consistent with stable, well-defined allele frequencies came from ABO and MN blood group surveys in the 1920s–1950s. Because ABO and MN genotypes can be read directly and unambiguously from a serological test (agglutination reactions with anti-A, anti-B, anti-M, anti-N sera), researchers such as Felix Bernstein (1924) could genotype thousands of individuals, tabulate genotype counts exactly as in the worked example above, and compute allele frequencies for A, B, and O. Bernstein's key contribution was showing that the ABO locus has three alleles at one locus (not two independent loci, as earlier workers proposed) — a conclusion he reached precisely by demonstrating that allele-frequency arithmetic based on a single three-allele locus fit the observed genotype counts, while a two-locus model did not. This is a direct historical example of allele-frequency calculation being used as the discriminating tool between competing genetic hypotheses, not merely as a descriptive summary.

Modern confirmation at the DNA level. Contemporary evidence that allele frequency is a real, measurable, biologically consequential quantity comes from genome-wide association studies (GWAS) and databases such as gnomAD, which report allele frequencies for millions of genetic variants across hundreds of thousands of sequenced genomes. These frequencies are computed by exactly the counting method shown above, scaled up: for the CFTR ΔF508 allele, population surveys of cystic fibrosis carrier screening data show a carrier (heterozygote) frequency of about 1 in 25 in populations of Northern European descent, from which q ≈ 0.02 and the homozygote (affected) frequency is predicted — and empirically observed — at roughly 1 in 2,500 to 1 in 3,500 births in that population, depending on the specific cohort. The consistency between counted carrier frequencies and observed disease incidence across independent studies is itself strong evidence that treating alleles as countable, stably-transmitted units correctly describes the underlying biology of inheritance in populations.

Example 1 — Sickle-cell allele frequency in a malaria-endemic population. A population genetics field study samples 600 individuals in a West African region with a long history of Plasmodium falciparum malaria, and genotypes each person at the β-globin locus for the sickle-cell allele (Hb^S) versus the normal allele (Hb^A). Results:

Total individuals N = 384 + 192 + 24 = 600. ✓ Step 1 — genotype frequencies:

• f(Hb^A Hb^A) = 384/600 = 0.64
• f(Hb^A Hb^S) = 192/600 = 0.32
• f(Hb^S Hb^S) = 24/600 = 0.04

(Sum = 0.64 + 0.32 + 0.04 = 1.00 ✓) Step 2 — allele frequencies, using the shortcut formula:

Check: 0.80 + 0.20 = 1.00 ✓ So in this population, the sickle-cell allele frequency is q = 0.20 — a strikingly high value compared to populations without historical malaria exposure, where q is typically well under 0.01. This large discrepancy is itself a clue (developed fully elsewhere) that heterozygote carriers are protected against malaria mortality, keeping the Hb^S allele at unusually high frequency despite the severe fitness cost of the homozygous sickle-cell disease genotype. Note that this calculation required no assumption about mating patterns or equilibrium — it is pure bookkeeping from genotype counts, exactly as in the mechanism section.

Example 2 — Distinguishing carrier frequency from allele frequency in clinical screening. A genetic counselor reviews a regional newborn screening report for classical phenylketonuria (PKU), an autosomal recessive metabolic disorder. The report states that 1 in 10,000 newborns in the sampled population is born with PKU (genotype aa), and asks a trainee to estimate (i) the allele frequency of the PKU allele and (ii) the expected carrier (Aa) frequency, assuming the population's genotype frequencies follow the Hardy-Weinberg proportions (an assumption to be justified elsewhere, but commonly used as a working approximation for rare recessive alleles). Given: f(aa) = 1/10,000 = 0.0001 Under Hardy-Weinberg genotype proportions, f(aa) = q², so:

Then p = 1 − q = 1 − 0.01 = 0.99 Expected carrier frequency, f(Aa) = 2pq = 2 × 0.99 × 0.01 = 0.0198 ≈ 0.02, i.e., about 1 in 50 individuals is a carrier. The pedagogically important point here is the contrast with Example 1: this calculation runs the arithmetic in the opposite direction, starting from a genotype frequency and an equilibrium assumption to infer allele frequency, rather than starting from raw genotype counts. A student must recognize which direction is licensed by the data: Example 1 needed no assumptions because full genotype counts were available directly; Example 2 needed the Hardy-Weinberg assumption because only the rarest genotype's frequency was known, and the heterozygote frequency (1/50, far larger than 1/10,000) can be dramatically undercounted by casual inspection — this is precisely why silent recessive carriers are so much more common in a population than the visibly affected homozygotes, a fact of major importance in genetic counseling.

Wilhelm Weinberg was a German physician practicing obstetrics and gynecology in Stuttgart, not a professional academic geneticist — he saw thousands of families through his medical practice and used that data to think rigorously about heredity in real human populations rather than in laboratory crosses. Independently, and in the same year, Godfrey Harold Hardy, a pure mathematician at Cambridge with famously little interest in applied science ("I have never done anything useful," he later wrote of his own work), derived the identical algebraic result almost as an aside, reportedly prompted by a colleague's genetics question over dinner and published as a short letter to the journal Science in 1908. The two men never collaborated and worked in completely different scientific cultures — Weinberg immersed in clinical human genetics and twin studies, Hardy immersed in number theory — yet arrived at the same simple but foundational algebra connecting allele frequency to genotype frequency under random mating. Weinberg's priority was largely unrecognized outside Germany for decades because his paper was published in German in a regional natural-history journal, and it was only in the mid-twentieth century that English-language geneticists (notably Curt Stern) properly credited Weinberg alongside Hardy, giving the principle its now-standard joint name.

Carrier screening programs for recessive genetic diseases — cystic fibrosis, Tay-Sachs disease, spinal muscular atrophy — report population allele frequencies to prospective parents so they can interpret personal genetic test results against a meaningful background rate, exactly the arithmetic worked through in Example 2. Forensic DNA profiling relies on measured allele frequencies at short tandem repeat (STR) loci across reference populations to calculate the probability that a DNA match occurred by chance, a number juries and courts depend on being both correct and defensible. Conservation geneticists estimate allele frequencies from small surviving populations of endangered species (cheetahs, black-footed ferrets, Tasmanian devils) to quantify loss of genetic diversity and guide captive breeding programs that try to preserve rare alleles before genetic drift eliminates them entirely. Plant and livestock breeders track allele frequencies for disease-resistance and yield-related loci across breeding generations to measure whether artificial selection programs are actually shifting the gene pool in the intended direction, using precisely the counting formulas developed here. Public health genomics databases such as gnomAD and the 1000 Genomes Project publish allele frequencies for millions of human genetic variants stratified by ancestry group, letting clinicians judge whether a variant found in a patient is a common, likely benign polymorphism or a rare variant more likely to be disease-causing — a judgment that is only possible because allele frequency can be measured with the exact method taught in this article.

Ask the tutor: How does the Hardy-Weinberg Principle use these allele frequencies to predict expected genotype frequencies, and how do we test whether a real population actually fits that prediction?