Allele Frequency Calculator

An allele frequency is a property of a gene pool rather than of any individual. In a diploid species each of N sampled individuals carries two alleles at a locus, giving 2N alleles in total, so the frequency of A is the count of A copies divided by 2N. Homozygotes contribute two copies each and heterozygotes one, which is why the heterozygote count appears in both the numerator for p and the numerator for q. Since every allele is either A or a, p plus q must equal exactly 1, the fastest arithmetic check on a calculation. The method used here is direct gene counting, and it depends on one condition that is easy to overlook: heterozygotes must be identifiable. That holds for codominant markers such as blood groups, allozymes and SNP genotypes, but not for a fully dominant trait, where AA and Aa are indistinguishable and q can only be inferred as the square root of the recessive phenotype frequency, an estimate that already assumes equilibrium and cannot then be used to test for it. Sampling precision follows the standard error of a proportion, approximately the square root of pq divided by 2N. At p equal to 0.5 a sample of 100 individuals gives a standard error near 0.035, so frequencies from small samples wobble considerably and should not be compared across populations without confidence intervals. The chi-square comparison against Hardy-Weinberg expectations carries one degree of freedom with a critical value of 3.841, and becomes unreliable when any expected genotype count falls below about 5. One frequent trap is pooling: combining two genetically distinct subpopulations produces an apparent deficit of heterozygotes even when each is internally at equilibrium, an artifact known as the Wahlund effect.

The vocabulary came before the arithmetic. Wilhelm Johannsen coined the terms gene, genotype and phenotype in 1909, shortly after Mendel's work was rediscovered. At that time British biology was split between Mendelians led by William Bateson, who saw inheritance as discrete factors, and biometricians around Karl Pearson and Walter Weldon, who studied continuously varying traits and doubted discrete factors could explain them. Ronald Fisher reconciled the two camps in his 1918 paper on the correlation between relatives on the supposition of Mendelian inheritance, showing that many small Mendelian effects sum to continuous variation. Over the 1920s and 1930s Fisher, J. B. S. Haldane and Sewall Wright built population genetics into a mathematical discipline in which evolution is defined precisely as change in allele frequency. Counting alleles empirically required markers where every genotype was visible. The MN blood group, described by Karl Landsteiner and Philip Levine in 1927, was among the first human loci where all three genotypes could be distinguished serologically, which is why MN data appear in so many classic worked examples including the one on this page. Scale arrived with protein electrophoresis in 1966, when Richard Lewontin and John Hubby working in Drosophila and Harry Harris working in humans independently revealed far more polymorphism than theory had anticipated. DNA genotyping and then large sequencing catalogues, notably the 1000 Genomes Project running from 2008, turned allele frequencies into routine reference data now used in clinical variant interpretation, where a high population frequency is treated as evidence against pathogenicity.