112 of the most-used terms in clinical and molecular genetics, each explained in two or three plain-English sentences. Skim the section chips below or search in-page with Ctrl+F.
The complete set of DNA you carry - every instruction for building and running the body. The human genome is about 3 billion DNA letters, held inside almost every nucleated cell.
Gene
A stretch of DNA that tells cells how to make something specific - usually a protein. Humans have around 20,000 genes, and small differences between people's genes explain a lot of what makes each of us unique.
Chromosome
A tightly packaged bundle of DNA and protein. Humans carry 23 pairs (46 total) in each cell - 22 pairs of autosomes plus one pair of sex chromosomes.
Allele
One of the possible versions of a given gene. You carry two alleles of most genes - one inherited from each parent - and the combination shapes traits from eye colour to disease susceptibility.
Locus
The precise street address of a gene or variant on a chromosome. Every gene has a fixed locus, so labs always know where to find it - for example, BRCA1 sits at locus 17q21.31.
Nucleotide
The basic "letter" of DNA. There are four nucleotides (A, C, G and T), and the order they appear in along a DNA strand is the genetic code.
Base pair
Two nucleotides bound across the DNA double helix. A always pairs with T, and C with G - like the teeth of a zip. Genome sizes are measured in base pairs (bp), kilobases (kb) or megabases (Mb).
Exon
The coding bits of a gene that are actually read to make a protein. Most human genes are a mix of exons (kept) and introns (trimmed out) along their length.
Intron
The non-coding spacer sections inside a gene. The cell splices them out before the gene is translated into protein, keeping only the exons.
UTR
The untranslated regions flanking a gene's coding sequence. The 5′ UTR sits before the coding part and the 3′ UTR sits after; both help control when, where and how much of the protein is made.
Codon
A triplet of three DNA letters that codes for one amino acid. The 64 possible codons (plus stop signals) map onto the 20 amino acids cells use to build proteins.
Amino acid
The molecular building block of proteins. Cells string together ~20 different amino acids in the exact order specified by a gene to produce each protein in your body.
Open reading frame
A stretch of DNA that, in principle, could be translated into protein - it starts with a start codon and ends with a stop codon. Finding open reading frames (ORFs) is one way genes are identified in newly-sequenced DNA.
Promoter / enhancer
DNA control switches for genes. The promoter sits immediately upstream of a gene and is essential for turning it on; enhancers can be far away but reach over in 3D space to boost the gene's expression.
Chromosome structure
Karyotype
A portrait of all 46 of your chromosomes, arranged in size order. Karyotyping spots big changes - an extra chromosome, a missing piece, a swap - and is still the standard first test for chromosomal disorders. Written as '46,XX' for females or '46,XY' for males.
Centromere
The pinched "waist" where the two arms of a chromosome meet. Cellular machinery clips onto the centromere during cell division to pull the chromosomes apart correctly.
Telomere
Protective caps on the ends of each chromosome - a bit like the aglets on shoelaces. Telomeres shorten a little each time a cell divides and are tied to cellular ageing.
p-arm and q-arm
Every chromosome has a short arm ('p', from the French petit) and a long arm ('q'). Addresses like 17q21 mean 'chromosome 17, long arm, region 2, band 1'.
Metacentric
A chromosome with its centromere roughly in the middle, giving arms of similar length. Chromosomes 1, 3, 19 and 20 are broadly metacentric.
Submetacentric
A chromosome with an off-centre centromere, producing one arm that is visibly shorter than the other. Most human chromosomes - including 4 through 12 - fall into this category.
Acrocentric
A chromosome with its centromere close to one end. In humans these are 13, 14, 15, 21 and 22 - each carrying distinctive "satellite" tips that house the ribosomal RNA genes.
Telocentric
A chromosome with its centromere right at the very tip. Humans don't have any telocentric chromosomes naturally - they turn up in some other species.
Sex chromosomes
The X and Y chromosomes that help determine biological sex. Females typically carry two X chromosomes and males typically carry one X and one Y; the remaining 22 pairs of chromosomes (called autosomes) are shared by both sexes.
Inheritance patterns
Autosomal dominant
A single copy of the variant - inherited from either parent - is enough to cause the condition. People with the variant typically have a 50% chance of passing it on to each child.
Autosomal recessive
Both copies of the gene (one from each parent) must carry the variant for the condition to appear. Carriers with only one affected copy are usually healthy but can still pass the variant on to their children.
X-linked dominant
The gene sits on the X chromosome and one affected copy causes the condition. Affected fathers pass it to all their daughters but none of their sons; affected mothers have a 50% chance with each child.
X-linked recessive
The gene sits on the X chromosome, and the condition typically affects males more severely because they have only one X chromosome. Female carriers usually have milder or no symptoms and can pass the variant to sons or daughters.
Y-linked
Passed from father to all sons via the Y chromosome. Only a handful of medical conditions follow this pattern because the Y chromosome carries relatively few genes.
Mitochondrial
Inherited exclusively from the mother via the mitochondrial DNA inside her egg cells. Affected mothers pass the variant to every child; affected fathers pass it to none.
De novo
A brand-new variant that arose in the egg, sperm or very early embryo rather than being inherited from either parent. Many severe paediatric conditions are caused by de novo variants, and the recurrence risk for future siblings is usually low.
Penetrance
The proportion of people with a variant who actually develop the condition. BRCA1, for instance, has high but incomplete penetrance - most but not all carriers develop breast or ovarian cancer during their lifetime.
Expressivity
How much the same variant's effects vary between people. Two siblings who inherit an identical variant can still end up with very different severities - that's variable expressivity in action.
Anticipation
When a condition tends to appear earlier or more severely with each successive generation. Anticipation is characteristic of repeat-expansion disorders like Huntington disease and myotonic dystrophy.
Compound heterozygous
Carrying two different variants in the same gene - one on the maternal copy and one on the paternal. Common in recessive conditions where each parent contributes a different change, for example CFTR variants in cystic fibrosis.
Homozygous
Carrying the same version of a gene on both chromosomes - two matching alleles. Being homozygous for a recessive variant is the classic set-up for autosomal recessive conditions.
Heterozygous
Carrying two different versions of a gene - one from each parent. This is the normal state for most gene-to-gene variation across the population.
Hemizygous
Having only one copy of a gene with no paired partner. Males are hemizygous for nearly every gene on the X chromosome, which is why X-linked recessive conditions affect them more often.
Carrier
Someone who carries a single copy of a recessive variant without showing the condition. When two carriers of the same recessive condition have a child together, there's a 1-in-4 chance the child will be affected.
Variants and classification
Variant
Any difference from a reference DNA sequence - the modern preferred term. 'Variant' is deliberately neutral: most variants are harmless, some increase disease risk, and a few cause disease directly.
Mutation
An older word for a DNA change, now often reserved for variants known to cause disease. Clinical reports increasingly use 'variant' plus an ACMG classification to avoid implying pathogenicity by default.
SNV
A single nucleotide variant - a change at just one DNA letter. SNVs are the most common kind of genetic variation between people.
SNP
A 'single nucleotide polymorphism' - essentially a SNV common enough (>1%) to count as natural population variation rather than a rare change. SNPs underpin genome-wide association studies and polygenic risk scores.
Indel
A small insertion or deletion of DNA - typically just a few letters. Even tiny indels can shift a gene's reading frame and disable the protein.
CNV
A copy number variant - a larger stretch of DNA that is gained (duplicated) or lost (deleted). CNVs can delete whole genes at once and cause conditions like DiGeorge syndrome or some autism risk variants.
Structural variant
A large-scale rearrangement of DNA - inversions, translocations or big CNVs. Structural variants are harder to detect than small variants and often need specialised sequencing.
Germline variant
A variant present in the reproductive cells, so it's heritable and present in every cell of the person carrying it. All the inherited conditions in this library are germline.
Somatic variant
A variant acquired during life - only in certain tissues, not passed on to children. Most cancer driver mutations are somatic variants that built up in a specific tissue over time.
Mosaicism
A mix of cells with different genotypes in one person - the result of a variant that arose after the embryo started developing. Mosaic variants can cause segmental conditions where only part of the body is affected.
ACMG classification
The five-tier scheme clinical labs use to interpret variants: Pathogenic, Likely pathogenic, Variant of uncertain significance (VUS), Likely benign and Benign. Only pathogenic and likely pathogenic variants are reported as causing disease.
VUS
A 'variant of uncertain significance' - the middle tier of the ACMG scheme. There isn't enough evidence yet to call it harmful or benign, so labs flag it for watch-and-wait until more data accumulate.
HGVS nomenclature
The international naming convention for variants - e.g. 'c.1521_1523delCTT' (DNA change) or 'p.Phe508del' (protein change). It pins down exactly which variant a report is describing.
How variants affect proteins
Missense
A DNA change that swaps one amino acid for a different one. Some missense variants have no effect; others subtly or dramatically change how the protein folds or works.
Nonsense
A DNA change that creates a premature stop codon, chopping the protein short. The truncated protein is usually non-functional - or destroyed by the cell's quality-control systems.
Silent
A DNA change that doesn't alter which amino acid is encoded - also called synonymous. Most silent variants are harmless, but a minority still affect splicing or translation efficiency.
Frameshift
An insertion or deletion that shifts the genetic reading frame by one or two letters. After the shift, every downstream codon is garbled, almost always producing a non-functional protein.
Splice-site variant
A variant at the boundary between an exon and intron that disrupts the cell's splicing machinery. The result is usually a protein missing an exon or containing intronic sequence it shouldn't.
Loss of function
A variant that reduces or eliminates the protein's activity - the most common way variants cause recessive disease.
Gain of function
A variant that gives the protein a new or heightened activity. Much rarer than loss-of-function, but important in some dominant conditions like achondroplasia (constitutively active FGFR3).
Haploinsufficiency
When one working copy of a gene is not enough to sustain normal function, so losing one copy causes disease. A common mechanism behind autosomal dominant conditions.
Dominant-negative
A variant where the faulty protein gums up the works of the healthy copy - producing disease even though one good allele remains. Characteristic of many collagen-gene conditions.
Null allele
A variant that produces no functional gene product at all - a complete knockout of that gene copy.
Cancer genetics
Oncogene
A gene whose activation drives cancer growth. Classic examples are RAS, MYC and HER2, which typically contribute to cancer when gained or over-activated, not when lost.
Proto-oncogene
A normal, healthy gene that - if altered in the right way - can become an oncogene. The vast majority of proto-oncogenes behave perfectly well until a specific activating change occurs.
Tumour-suppressor gene
A gene that restrains cell growth or triggers death in damaged cells. When both copies of a tumour-suppressor stop working (think TP53, BRCA1/2, RB1, APC), cells can grow unchecked.
Knudson two-hit hypothesis
The idea that both copies of a tumour-suppressor gene need to be knocked out before cancer can develop. Inherited tumour-predisposition syndromes start with one 'hit' already in place, so only a second somatic hit is needed - explaining why cancers occur earlier in these families.
Driver mutation
A variant that actively contributes to the growth of a tumour. A tumour typically has a handful of driver mutations alongside many passenger mutations.
Passenger mutation
A variant present in a tumour but not actually contributing to its growth - a genetic bystander that came along for the ride. Modern cancer genomics focuses on separating drivers from passengers.
Loss of heterozygosity
When one of the two parental alleles at a locus is lost - commonly seen in tumours as a way of knocking out the surviving healthy copy of a tumour-suppressor gene.
Microsatellite instability
A fingerprint of faulty mismatch-repair machinery, showing up as unstable short repeat sequences in a tumour's DNA. A classic feature of Lynch-syndrome-related cancers, and a marker that the tumour may respond to immunotherapy.
Tumour mutational burden
The total number of mutations found in a tumour genome. High mutational burden correlates with response to immune checkpoint inhibitors in several cancer types.
Population and quantitative genetics
Allele frequency
The proportion of a given allele in a population. Rare disease variants often have allele frequencies below 1 in 10,000; common benign variants are seen in >5% of the population.
Minor allele frequency
Shorthand (MAF) for the frequency of the less common of the two alleles at a site. Clinical labs use MAF to help triage new variants - a variant seen at 10% in healthy databases is very unlikely to cause a rare severe disease.
Hardy-Weinberg equilibrium
The mathematical baseline that predicts allele frequencies in an idealised population with no selection, migration or genetic drift. Real populations deviate from Hardy-Weinberg, but it remains the reference point for spotting those deviations.
Founder effect
The shift in allele frequencies that happens when a small group founds a new population. Well-known examples are the elevated BRCA1/2 founder variants in Ashkenazi Jewish communities and the CFTR F508del variant in Northern Europeans.
Linkage disequilibrium
The tendency for certain alleles at nearby sites to travel together down the generations rather than being reshuffled. LD is what lets genome-wide association studies identify risk regions instead of having to test every single variant.
GWAS
A genome-wide association study - scanning the genomes of large groups of people to find variants statistically associated with traits or disease. Most GWAS hits have small individual effects, which is why polygenic risk scores combine many of them.
Polygenic risk score
A single number that sums up the effects of many common variants, giving an estimate of genetic risk for a complex trait like coronary artery disease. A PRS isn't a diagnosis - it's a nudge on the risk scale, most useful when combined with clinical and lifestyle factors.
Heritability
The share of variation in a trait across a population that can be attributed to genetic variation. High heritability of a trait does not mean an individual outcome is pre-determined - environment, lifestyle and chance still matter.
Methods and technologies
PCR
The polymerase chain reaction - a lab trick that copies a specific piece of DNA billions of times over. PCR underpins everything from COVID tests to fragment analysis in genetic labs.
Sanger sequencing
The original, high-accuracy method of reading DNA one letter at a time. Still used today to confirm single-variant findings, but too slow to sequence whole genomes on its own.
Next-generation sequencing
Massively parallel sequencing technologies (NGS) that read millions of DNA fragments simultaneously. NGS powers modern panel tests, exome sequencing and whole-genome sequencing at a fraction of the cost of Sanger.
Targeted panel
An NGS test that only sequences a curated set of genes - usually those relevant to a clinical question. A carrier-screening panel, for instance, might cover 600 recessive-disease genes.
Whole exome sequencing
Sequencing the ~2% of the genome that codes for protein (the exome). Because most known disease-causing variants sit in coding regions, exome sequencing is a cost-effective first-line diagnostic test.
Whole genome sequencing
Sequencing the entire 3-billion-letter genome - coding regions, introns, regulatory DNA and everything in between. Whole-genome sequencing catches structural variants and regulatory changes that exome sequencing can miss.
Chromosomal microarray
A genome-wide scan for copy number variants and common SNPs, now standard first-line for unexplained developmental delay or multiple congenital anomalies. Higher resolution than a karyotype but still misses single-letter variants.
FISH
Fluorescence in situ hybridisation - a technique that lights up specific DNA sequences on microscope slides of chromosomes. Useful for confirming suspected deletions, duplications or translocations spotted on karyotype.
MLPA
Multiplex ligation-dependent probe amplification - a targeted way to detect deletions and duplications of specific exons. Often used when a microarray flags a region of interest or when a gene has known deletion-prone regions.
Karyotyping
The microscope-based analysis of a full set of chromosomes, looking for changes in number or large-scale structure. Still the test of choice when a full chromosomal picture is needed, for instance in infertility work-up.
Long-read sequencing
Newer sequencing technologies (PacBio, Oxford Nanopore) that read very long DNA fragments. Long reads resolve structural variants and repeat expansions that short-read NGS struggles with.
Prenatal and reproductive genetics
Aneuploidy
An abnormal chromosome number - the most familiar example being trisomy 21 (Down syndrome). Aneuploidy is one of the commonest causes of miscarriage and of chromosomal conditions detected prenatally.
Trisomy
Having three copies of a chromosome instead of the usual two (e.g. trisomy 21, 18 or 13). Most trisomies are incompatible with life; those compatible with life cause recognisable chromosomal syndromes.
Monosomy
Having only one copy of a chromosome when there should be two. Most monosomies are lethal very early in development; the main survivable exception is monosomy X (Turner syndrome).
Polyploidy
An extra complete set of chromosomes - for example, triploidy (69 chromosomes total instead of 46). Human polyploidies are almost always incompatible with long survival.
NIPT
Non-invasive prenatal testing - analysing placental DNA fragments circulating in the mother's blood from around 10 weeks of pregnancy. NIPT gives a highly sensitive screen for the common trisomies without any procedural risk to the pregnancy.
CVS
Chorionic villus sampling - a diagnostic test that takes a small placental biopsy at 11–14 weeks of pregnancy. CVS gives a definitive answer but carries a small (~0.2%) procedure-related miscarriage risk.
Amniocentesis
A diagnostic test that samples amniotic fluid from 15 weeks of pregnancy onwards. Like CVS, amniocentesis gives a definitive diagnosis with a small procedure-related miscarriage risk.
Preimplantation genetic testing
Testing embryos created through IVF before they are transferred. Variants include PGT-A (for chromosomal aneuploidy), PGT-M (for a specific inherited monogenic condition) and PGT-SR (for structural rearrangements).
Genomic imprinting
A phenomenon where the two parental copies of certain genes behave differently depending on which parent they came from. Imprinting disorders - like Prader-Willi and Angelman syndromes - arise when the parent-of-origin-specific gene is disrupted.
Uniparental disomy
When both copies of a chromosome come from the same parent, rather than one from each. UPD usually goes unnoticed, but if it affects an imprinted region it can mimic a deletion disorder.
Confined placental mosaicism
Mosaicism (a mix of cell populations) that is limited to the placenta and not present in the fetus. CPM is an important reason prenatal screening results need to be interpreted carefully and sometimes followed up with diagnostic testing.
Epigenetics and gene regulation
Epigenetics
Heritable changes in gene activity that don't alter the DNA sequence itself. Epigenetic marks are the body's volume control for genes - turning them up or down depending on cell type, developmental stage or environmental cues.
DNA methylation
The chemical attachment of methyl groups to DNA - one of the main ways cells silence genes they no longer need. Methylation patterns are read by clinical labs to diagnose imprinting disorders and some cancers.
Histone modification
Chemical tags added to the histone proteins that DNA wraps around. These tags loosen or tighten the local DNA packaging, and in doing so control which genes are available to be transcribed.
X-inactivation
The process by which each cell in a female randomly silences one of her two X chromosomes early in development (sometimes called Lyonisation). X-inactivation explains why female carriers of X-linked recessive conditions can show mild or patchy features.
Non-coding RNA
RNA that doesn't code for protein but plays regulatory roles - including microRNAs (miRNAs) and long non-coding RNAs (lncRNAs). Increasingly implicated in disease, they are a fast-moving area of research.
Key databases and resources
OMIM
Online Mendelian Inheritance in Man - an online catalogue of genes and inherited conditions, linking each disorder to the genes and variants that cause it. The 'OMIM number' (e.g. #300040) is the standard way clinicians and researchers refer to a specific condition.
ClinVar
A public NCBI database of variants and their reported clinical interpretations. Labs worldwide contribute their findings, and ClinVar is where clinicians look up whether a variant has been classified before.
gnomAD
The Genome Aggregation Database - allele frequencies from well over a million exomes and genomes. gnomAD is the first stop for checking how common (or rare) a variant is across populations.
RefSeq, Ensembl and UCSC
The three big reference resources for genome sequence and annotation. RefSeq (NCBI) and Ensembl are the authoritative gene-model databases; the UCSC Genome Browser is the go-to visualisation tool for looking at genome regions in context.
HGNC
The HUGO Gene Nomenclature Committee - the body that assigns and curates official human gene symbols. HGNC is the reason every gene has a single, standardised symbol like BRCA1 instead of a scrum of old aliases.
LOVD and DECIPHER
Curated variant and phenotype databases. LOVD focuses on locus-specific variant repositories (typically per gene); DECIPHER focuses on rare structural variants and developmental phenotypes.
ACMG / AMP guidelines
The 2015 framework (Richards et al.) that labs use worldwide to classify variants into the five ACMG tiers. It defines the lines of evidence that count towards pathogenic or benign interpretations, and remains the reference standard.
This site uses cookies
We use cookies to measure how the site is used and to improve it. No
analytics or advertising cookies are set until you choose. See our
Privacy Policy and
Terms & Conditions.
Strictly necessary (always on)
Sign-in, your basket and security. The site cannot work without these.
Functional
Remembers choices you have made, such as your selected test options.
Analytics
Google Analytics, so we can see which pages help and which do not.
Marketing
Google Ads and Meta measure whether an advert led to a visit.
