Haplotypes, Genotypes, and Phase at Two Loci

Why Two Loci Require Two Levels of Description

At one genetic locus, a diploid individual carries two allele copies, one on each homologous chromosome. With two loci, the individual still has two copies at each locus, but now the alleles are also arranged along the two chromosomes. A genotype recorded separately at each locus tells us which alleles are present there; it may not tell us how the alleles at the two loci are paired on the chromosomes.

That distinction is the starting point for describing genetic variation across loci. A haplotype describes the allele combination on one chromosome, while a multilocus genotype describes the alleles carried by an individual at both loci. Phase specifies which alleles at the two loci lie together on each of the individual's two homologous chromosomes.

Definitions: A locus is a specified position or region in the genome. An allele is one of the alternative forms considered at that locus. A haplotype at two loci is the ordered pair of alleles carried together on one chromosome, with the locus order fixed. A genotype at a locus is the pair of alleles an individual carries at that locus. A two-locus multilocus genotype gives the individual's genotype at each of the two loci. Phase is the assignment of those alleles to the two chromosome copies.

Throughout this tutorial, assume an autosomal region in an ordinary diploid individual, with two specified loci and two possible alleles at each locus. Write the alleles at the first locus as \(A\) and \(a\), and those at the second locus as \(B\) and \(b\). The order is always first locus, then second locus. Thus \(AB\) means an \(A\) allele and a \(B\) allele on the same chromosome; it does not mean an individual has genotype \(A/A\) and \(B/B\).

At a single locus, the three possible genotypes are \(A/A\), \(A/a\), and \(a/a\), or, at the second locus, \(B/B\), \(B/b\), and \(b/b\). The slash separates the two alleles carried by an individual at that locus. The order on either side of the slash is not meaningful: \(A/a\) and \(a/A\) describe the same genotype.

From Chromosomes to a Two-Locus Genotype

A diploid individual has two homologous chromosome copies across the region. Each chromosome carries one allele at the first locus and one at the second. For example, the chromosome pair \(AB\) and \(ab\) contains an \(A\) and an \(a\) at the first locus, and a \(B\) and a \(b\) at the second. If the data report the two loci separately, the observed multilocus genotype is \((A/a,\ B/b)\).

The parentheses here group the two single-locus genotypes in locus order. They do not specify which allele at the first locus shares a chromosome with which allele at the second locus. That extra information is phase. To describe a phased individual, write the two chromosome haplotypes together, such as \(AB/ab\). The slash now separates chromosome copies; its meaning is determined by the notation in use.

Worked Example: Translating a Phased Chromosome Pair

Suppose a hypothetical diploid individual has the two chromosome haplotypes \(AB\) and \(aB\), with the loci written in the fixed order first then second. Find the genotype at each locus and the two-locus multilocus genotype.

First locus. Read the first character of each chromosome haplotype: one chromosome carries \(A\), and the other carries \(a\). The first-locus genotype is \(A/a\).

Second locus. Read the second character of each haplotype: both chromosomes carry \(B\). The second-locus genotype is \(B/B\).

Therefore, the multilocus genotype is \((A/a,\ B/B)\). In this example the chromosome pair is \(AB/aB\), and the separate locus genotypes are obtained by reading across the two haplotypes. Because both chromosomes have \(B\) at the second locus, there is no alternative pairing of the observed alleles that changes the chromosome-level description.

The same translation works in the reverse direction, but it may not give a unique answer. If the single-locus genotypes are \(A/a\) and \(B/b\), the individual has one copy of each allele at each locus. There are two ways to place those alleles on the two chromosomes: \(AB/ab\), or \(Ab/aB\). Both arrangements produce the same alleles at each locus when the chromosomes are read separately.

Key distinction: A multilocus genotype records which alleles an individual carries at each locus. A phased haplotype pair records how those alleles are joined along the two chromosomes. Different haplotype pairs can therefore be consistent with the same multilocus genotype.

Phase: What Is Known and What Is Not

For the double-heterozygous multilocus genotype \((A/a,\ B/b)\), the two compatible phase configurations are \(AB/ab\) and \(Ab/aB\). These are distinct pairings, not different spellings of the same pairing. In \(AB/ab\), \(A\) is on the same chromosome as \(B\); in \(Ab/aB\), \(A\) is on the same chromosome as \(b\).

When an assay reports only the genotype at each locus, it identifies \((A/a,\ B/b)\) but does not, by that information alone, identify which of the two configurations is present. Phase may be established by data that retain chromosome-level information, such as a suitable sequencing read or family transmission information. The important point here is not how to estimate or resolve phase, but what information a phase statement contains.

In other genotype combinations, phase may be determined directly by the alleles. For example, if the first-locus genotype is \(A/A\), every chromosome has \(A\) at that locus. Pairing the alleles at the second locus with \(A\) cannot create two distinct configurations. The same logic applies if the second-locus genotype is homozygous. If both loci are heterozygous, however, each locus by itself leaves open which allele is paired with which allele at the other locus.

Worked Example: Two Phases, One Unphased Genotype

A hypothetical diploid sample is called \(A/a\) at the first locus and \(B/b\) at the second. The assay reports no chromosome-specific information. List the compatible phase configurations and explain why the call does not distinguish them.

There are two alleles at each locus, one on each chromosome. Put \(A\) on one chromosome. The remaining allele at the first locus, \(a\), must be on the other. The \(B\) allele can be on the chromosome carrying \(A\), giving \(AB/ab\), or it can be on the chromosome carrying \(a\), giving \(Ab/aB\).

Checking the first configuration: \(AB/ab\) gives \(A/a\) at the first locus and \(B/b\) at the second. Checking the second: \(Ab/aB\) also gives \(A/a\) and \(B/b\). Thus both yield the reported multilocus genotype, and the unphased call alone cannot select between them.

Interchanging the labels “first chromosome” and “second chromosome” does not create another configuration: \(AB/ab\) and \(ab/AB\) describe the same pair of haplotypes. The two genuine possibilities differ in which alleles are paired together, not in the arbitrary naming of the chromosomes.

Counting the Genotype Classes and Phase Possibilities

At each biallelic locus there are three unordered genotypes: homozygous for the first allele, heterozygous, or homozygous for the second allele. A two-locus multilocus genotype is a choice of one genotype at each locus. Therefore, the number of possible unphased multilocus genotypes is the product of the number at each locus:

3 \times 3 = 9

This count concerns genotype classes, not the number of chromosome-level haplotype pairs. To see the phase distinction, consider the one genotype class that is heterozygous at both loci. Assigning which first-locus allele is on the first chromosome fixes the other first-locus allele on the second. The second-locus alleles can then be paired in two ways. Relabeling the two chromosomes swaps both haplotypes and leaves the same pair, so there are exactly two distinct phase configurations.

When either locus is homozygous, switching which chromosome carries an allele at that locus cannot produce a different pairing. For instance, \((A/A,\ B/b)\) can only be represented as \(AB/Ab\), up to swapping chromosome labels. Together, these facts explain why phase ambiguity is specific to the double-heterozygous genotype in the two-locus, biallelic setting.

Worked Example: A Genotype with Phase Determined by Homozygosity

A hypothetical individual has multilocus genotype \((A/A,\ B/b)\). Determine the chromosome haplotypes and show why there is no second distinct phase configuration.

At the first locus, both chromosome copies must carry \(A\). At the second locus, one chromosome carries \(B\) and the other carries \(b\). Pairing these alleles gives \(AB/Ab\).

The alternative-looking arrangement \(Ab/AB\) merely swaps the chromosome labels and describes the same chromosome pair. There is no arrangement such as \(AB/ab\), because that would put \(a\) on one chromosome and contradict the observed \(A/A\) genotype. Thus this multilocus genotype uniquely determines the haplotype pair, apart from arbitrary chromosome labeling.

A Short Justification for the Two-Configuration Result

The phase result can be checked by enumerating the pairings while respecting the genotype. For a double heterozygote, each chromosome must receive one of \(A,a\) and one of \(B,b\). Fix the first chromosome temporarily as the one carrying \(A\); this is only a bookkeeping choice, since chromosome labels have no intrinsic meaning. The second chromosome must then carry \(a\). There are two remaining choices for the first chromosome's second-locus allele:

\begin{aligned} A \text{ paired with } B &\quad\Longrightarrow\quad AB/ab,\\ A \text{ paired with } b &\quad\Longrightarrow\quad Ab/aB. \end{aligned}

The corresponding second chromosome is forced in each case because the individual has exactly one \(B\) and one \(b\). No third pairing is possible: placing \(B\) on both chromosomes would contradict \(B/b\), and placing \(b\) on both would also contradict \(B/b\). The enumeration is complete, and the two listed arrangements differ in their chromosome-level allele pairings.

Common Mistakes and Why the Distinction Matters

Treating a haplotype as an individual's genotype. \(AB\) is the allele combination on one chromosome, not the two-locus genotype of a diploid individual. For an individual, \(AB/ab\) names a pair of chromosome haplotypes.
Assuming the two single-locus genotypes reveal phase. The call \((A/a,\ B/b)\) gives the alleles present at both loci, but it is compatible with both \(AB/ab\) and \(Ab/aB\). A chromosome-level pairing requires phase information.
Counting chromosome labels as different phases. \(AB/ab\) and \(ab/AB\) represent the same pair, because swapping the names of the two homologues changes no biological information.
Confusing allele identity with chromosome location. Knowing that \(A\) and \(B\) are both present does not establish that they are on the same chromosome. The phase statement is precisely what records that pairing.

Keeping these descriptions separate matters whenever a question concerns combinations of alleles carried together on chromosomes. An unphased genotype is appropriate for reporting which alleles occur at each locus; a haplotype pair is needed to state the chromosome-level arrangement. Neither description should be substituted for the other without saying what information is available.

Key takeaway: At two biallelic loci, a diploid individual has two chromosome haplotypes and one multilocus genotype. The multilocus genotype records the two alleles at each locus; phase records how those alleles are paired across chromosomes. A double heterozygote has two compatible phase configurations when only its unphased genotype is known.

Check Your Understanding

Use the fixed locus order of first locus, then second locus.

What multilocus genotype results from the chromosome haplotypes \(Ab\) and \(aB\)?
List the two phase configurations compatible with the unphased genotype \((A/a,\ B/b)\).
Why do \(AB/ab\) and \(ab/AB\) not count as two distinct phase configurations?
Give the haplotype pair implied by \((a/a,\ B/b)\), ignoring the arbitrary order of chromosome labels.
How many possible unphased multilocus genotypes are there with two biallelic loci, and what simple count gives this number?