Population genetics describes genetic variation using frequencies rather than only individual genotypes.

For a diploid locus with alleles $A$ and $a$, let the genotype counts be

$$N_{AA},\quad N_{Aa},\quad N_{aa},$$

with total individuals

$$N=N_{AA}+N_{Aa}+N_{aa}.$$

The genotype frequencies are simply each count divided by $N$.

The population contains $2N$ allele copies at this autosomal locus. The frequency of allele $A$ is

$$p=\frac{2N_{AA}+N_{Aa}}{2N},$$

and the frequency of allele $a$ is

$$q=\frac{2N_{aa}+N_{Aa}}{2N}.$$

Because these are the only two alleles,

$$p+q=1.$$

Worked example

Suppose 100 individuals have genotypes

• 36 $AA$,
• 48 $Aa$,
• 16 $aa$.

Then

$$p=\frac{2(36)+48}{200}=\frac{120}{200}=0.60,$$

and

$$q=0.40.$$

The genotype frequencies are $0.36$, $0.48$ and $0.16$, while the allele frequencies are $0.60$ and $0.40$. These are different kinds of quantities.

Allele frequencies track the composition of the gene pool, the collection of allele copies present in the population. Evolution at a locus can therefore be detected by comparing allele frequencies across generations.