What is an allele?

An allele is one of two or more versions of a DNA sequence (a single base or a segment of bases) at a given genomic location, or locus. The locus is the position; the alleles are the different versions that can sit there. A person with two identical alleles at a locus is homozygous there, and a person with two different alleles is heterozygous.

A person inherits two alleles at each autosomal locus, one from each parent. The two alleles lie at the same position on a pair of homologous chromosomes, one maternal and one paternal. (X-linked loci in males, and mitochondrial DNA, are exceptions.)

Where the word comes from

Allele is a shortened form (adopted in 1931) of allelomorph (1902), meaning “alternative form of a gene”. The root is the Greek allel-, “one another”: the alleles are the mutual, reciprocal alternatives at a locus.

What sequencing tells us

DNA sequencing determines the order of nucleotides at a locus, and so reveals which alleles a person carries there.

Variant allele frequency (VAF) is the proportion of sequencing reads covering a position that carry the variant: variant reads divided by total read depth. It is a guide to the allelic configuration:

• A heterozygous germline variant is expected at about 50%.
• A homozygous variant is expected at about 100%.
• Lower values suggest that only some cells carry the variant, as in a tumour sample or mosaicism. For example, a heterozygous variant present in every tumour cell, in a sample that is 50% tumour cells (and assuming no copy-number change), would give a VAF of about 25%.
• These are expectations, not rules. Copy-number changes, allelic bias in the assay, contamination and sequencing errors can all shift a VAF away from its expected value.

VAF helps identify and classify variants. Deciding whether a variant is the causative mutation takes further interpretation, such as its predicted effect, its frequency in the population and whether it tracks with disease in the family.

Chromosomes and chromatids

The terms depend on where the cell is in its cycle.

• Before DNA replication (G1 phase): a chromosome is a single DNA molecule, which is a single chromatid. Chromosomes are present throughout the cell cycle, but at this stage they are too decondensed to be seen as separate structures under a light microscope.
• After replication (from the end of S phase): the cell has copied each chromosome, so each one now consists of two identical sister chromatids joined at the centromere. A typical human cell still has 46 chromosomes, but it now has 92 chromatids.
• When chromosomes become visible, the DNA has already replicated. Chromosomes condense during cell division, and the familiar X-shaped chromosome is a pair of sister chromatids.
• In anaphase, the sister chromatids separate, and each is then counted as a chromosome in its own right.

Both sister chromatids are copies of the same chromosome, so they come from the same parent. They are identical apart from rare copying errors. Do not confuse them with homologous chromosomes: the maternal and paternal chromosomes are separate homologues, and after replication each has its own pair of sister chromatids.

In mitosis the sister chromatids are pulled apart, so each daughter cell receives one copy of every chromosome. In meiosis I the homologous chromosomes separate, and in meiosis II the sister chromatids separate. During meiosis, crossing over exchanges segments between non-sister chromatids (one maternal and one paternal), so a chromatid in a gamete can carry stretches of DNA from both of a person’s parents.

A note on the level at which tests look

Different genetic tests examine DNA at very different scales, and the scale decides which kind of change they can find. Karyotyping sees whole chromosomes and changes of about 5 Mb or more. FISH typically detects deletions larger than about 200 kb, and chromosomal microarray reaches down to tens of kilobases or finer, depending on the platform. MLPA works at the level of exons and genes, and sequencing works at the level of individual bases, which is where alleles are read. A normal result at one level does not exclude a change at another.