Identical twins share very close to 100 percent of their DNA sequence, but “very close” is not the same as “perfectly identical.” Fraternal twins, by contrast, share roughly 50 percent of their inherited genetic variants, the same average as any pair of siblings. That textbook framing is a useful starting point, but the real picture has become more interesting as sequencing technology has improved. Somatic mutations, copy-number differences, epigenetic changes, and even a rare third category of twin all complicate the simple percentages.
Where the “100 Percent” Figure Comes From
Identical twins form when a single fertilized egg splits, so both embryos begin life with the same genome. Traditional twin studies treat this as a given: monozygotic twins share 100 percent of their genes, while dizygotic twins share about 50 percent.1PubMed Central. Twin Study Design That baseline holds for the vast majority of the genome. If you lined up the roughly three billion base pairs in one identical twin’s DNA against those of the other, almost every letter would match.
The exceptions are what make the story interesting. From the moment the embryo splits, each twin’s cells begin dividing independently. Every cell division carries a small chance of a copying error, and over the trillions of divisions that build a human body, some errors accumulate in one twin but not the other. These are called somatic mutations, and they are the primary reason identical twins are not genetically interchangeable. In one study of twins discordant for psychiatric disorders, researchers confirmed somatic mutations with allele fractions ranging from about 1 to 7 percent in the affected twin, while the same positions in the co-twin showed essentially zero signal.2PubMed Central. Identification of somatic mutations in monozygotic twins discordant for psychiatric disorders These are not inherited differences. They arise during development and can end up in only a fraction of one twin’s cells, but they are genuine DNA-level distinctions.
Copy Number Variations Add Another Layer
Beyond single-letter typos, twins can also differ in stretches of DNA that get duplicated or deleted. These structural changes, known as copy number variations, have been found within identical twin pairs regardless of whether the twins looked or acted alike. Genome-wide screening of 19 monozygotic twin pairs detected copy number differences in both phenotypically concordant and discordant pairs.3PubMed Central. Phenotypically concordant and discordant monozygotic twins display different DNA copy-number-variation profiles A separate study of discordant twins found de novo copy number duplications overlapping genes previously linked to conditions like autism spectrum disorder.4PubMed Central. De novo single-nucleotide and copy number variation in discordant monozygotic twins reveals disease-related genes
These differences are not trivial. In families with monozygotic twins discordant for schizophrenia, individuals carried between 35 and 65 copy number variants each, with roughly 10 percent of those being de novo, meaning they were not present in either parent. Of those de novo changes, about 70 percent arose during developmental cell division rather than during the parents’ reproductive process.5PubMed Central. Ontogenetic de novo copy number variations (CNVs) as a source of genetic individuality: studies on two families with MZD twins for schizophrenia In other words, most of the unique structural variants in each twin’s genome were generated after conception, during the building of their individual bodies.
Same Genome, Different Settings
Even where the DNA sequence itself is identical, the way it gets read can diverge. Chemical tags on DNA and the proteins that package it control which genes are active in which cells. These epigenetic marks are influenced by environment, diet, and random chance, and they shift over a lifetime. A landmark study of monozygotic twins found that young twins were epigenetically indistinguishable, but older twin pairs showed striking differences in DNA methylation and histone acetylation patterns across their genomes, with measurable effects on gene expression.6PubMed Central. Epigenetic differences arise during the lifetime of monozygotic twins
This drift is not always gradual. Tracking twins from birth to 18 months revealed that some pairs diverged in DNA methylation rapidly while others stayed stable or even converged. The sites that changed most were concentrated in genes involved in development, suggesting that epigenetic individuality begins forming in the earliest months of life.7PubMed Central. Longitudinal, genome-scale analysis of DNA methylation in twins from birth to 18 months of age reveals rapid epigenetic change in early life and pair-specific effects of discordance Remarkably, at certain spots on the genome, the methylation variation between identical co-twins was as large as the variation between unrelated people, pointing to a kind of universal epigenetic noise that genetics alone cannot explain.8PubMed Central. Equivalent DNA methylation variation between monozygotic co-twins and unrelated individuals reveals universal epigenetic inter-individual dissimilarity
Epigenetic changes do not alter the DNA sequence, so in the strictest molecular sense they are not differences in “shared DNA.” But they affect which genes are turned on and off, and they can be passed along when cells divide. For anyone interested in why identical twins sometimes develop different diseases, epigenetics is arguably as important as the sequence itself.
How Much DNA Do Fraternal Twins Share?
Fraternal twins come from two separate eggs fertilized by two separate sperm, making them genetically equivalent to any pair of siblings. The textbook figure is 50 percent of their segregating genetic variants, meaning the variants that differ between their parents.1PubMed Central. Twin Study Design In practice, the actual percentage for any given pair can range anywhere from somewhat below to somewhat above that average, depending on which chromosomal segments each twin happened to inherit. Two fraternal twins might share 40 percent or 60 percent at any given locus; 50 percent is just the expected value across the whole genome.
People sometimes assume fraternal twins must be more genetically similar than ordinary siblings because they develop in the womb at the same time, but shared prenatal timing does not change which chromosomes get dealt to each embryo. What it does change is the environment: fraternal twins share the same uterine conditions, maternal nutrition, and gestational exposures. That shared environment can make them more similar in certain traits than siblings born years apart, but the similarity is environmental, not genetic.
The Rare In-Between Category
In 2019, a case report described twins who were neither fully identical nor fully fraternal. These “sesquizygotic” twins shared all of their mother’s DNA but were chimerically sharing about 78 percent of their father’s genome.9PubMed. Molecular Support for Heterogonesis Resulting in Sesquizygotic Twinning The suspected mechanism involves a single egg being fertilized by two sperm, followed by division of the resulting cell mass into two embryos.10PubMed. The New Sesquizygotic Twins and More: Exotic Twin Types That puts them genetically somewhere between identical and fraternal, sharing more than the typical 50 percent but less than the near-100 percent of monozygotic twins.
Sesquizygotic twinning is extremely rare and had only been confirmed once at the time of publication. It is not clear how often it occurs undetected, since most twin-zygosity testing does not look for the mixed paternal chimerism that would reveal it. The case matters because it breaks the neat binary of “identical or fraternal” that most discussions rely on.
X-Chromosome Inactivation and Discordant Disease
Female identical twins share the same two X chromosomes, but which copy gets silenced in each cell is partly random. This process, called X-chromosome inactivation, happens early in development, and the pattern can differ between twins. One dramatic case involved monozygotic female twins born to a hemophilia A carrier mother. One twin had severe hemophilia with less than 1 percent of normal clotting factor activity, while the other was an asymptomatic carrier with 42 percent activity. The explanation: the affected twin’s cells had nonrandomly silenced the healthy paternal X chromosome, leaving the hemophilia-carrying maternal X active, while the other twin had a random inactivation pattern that left enough healthy copies active.11PubMed Central. Female monozygotic twins discordant for hemophilia A due to nonrandom X-chromosome inactivation
This is a situation where two people share identical DNA sequences but have profoundly different clinical outcomes because of how their cells chose to use that DNA. It is most relevant for X-linked conditions, but the broader principle applies: genetic identity at the sequence level does not guarantee functional identity at the cellular level.
Mitochondrial DNA in Twins
Human cells carry a small circular genome inside their mitochondria, inherited solely from the mother. Because both twins originate from the same fertilized egg, they start with the same mitochondrial DNA. Studies sequencing mitochondrial genomes in identical twin pairs have found that high-confidence variants are shared between co-twins.12PubMed Central. Twin mitochondrial sequence analysis Research into whether low-level mitochondrial sequence variation, known as heteroplasmy, differs between identical twins and could explain disease discordance has generally come up negative. Studies of twins discordant for schizophrenia found mitochondrial heteroplasmy levels to be nearly identical between co-twins, suggesting that the mitochondrial genome is divided fairly evenly when the embryo splits.13PubMed. mtDNA Heteroplasmy in Monozygotic Twins Discordant for Schizophrenia14PubMed. Mitochondrial DNA Copy Number and Heteroplasmy in Monozygotic Twins Discordant for Schizophrenia
Interestingly, the very stability of mitochondrial DNA between identical twins has turned into a forensic tool. When nuclear DNA is too similar to distinguish twins, mitochondrial heteroplasmy differences, where they exist, can sometimes provide the distinguishing signal.
How Forensic Science Tells Twins Apart
Standard forensic DNA profiling looks at a panel of short tandem repeat markers. Identical twins produce indistinguishable profiles on these tests, which has created real problems in criminal cases and paternity disputes. The workaround, developed in the last decade, relies on the somatic mutations described earlier. By sequencing twin DNA at extreme depth, researchers can find rare single-letter mutations present in one twin but not the other. In one paternity case, ultra-deep sequencing of sperm samples from two identical twins identified five unique mutations in the father that were also present in his child but absent from the uncle.15PubMed. Finding the needle in the haystack: differentiating “identical” twins in paternity testing and forensics by ultra-deep next generation sequencing
A criminal case report demonstrated a similar approach, combining whole-genome sequencing with targeted molecular analysis. The investigators found that mitochondrial DNA actually had higher potential for distinguishing the twins than nuclear DNA in that particular case, because one twin carried heteroplasmic variants the other did not.16PubMed. Identification of the perpetrator among identical twins using next-generation sequencing technology: A case report These methods are expensive and not yet routine, but they establish that the genetic differences between identical twins, while tiny, are real and detectable.
Blood Chimerism and Shared Cells
Twins who share a placenta can exchange blood stem cells through vascular connections, creating a situation where each twin carries a small population of cells with the other twin’s genome. This happens most often in monochorionic dizygotic twins, a rare arrangement where fraternal twins develop within a single placenta. In these cases, blood-based DNA testing can show markers from both twins, confirming chimerism through the shared vascular network.17PubMed. Confined blood chimerism in monochorionic dizygous (MCDZ) twins This blood chimerism can persist for years after birth.18PubMed. Blood Chimerism in Dizygotic Monochorionic Twins During 5 Years Observation
Even when a twin pregnancy ends with only one surviving baby, cells from the lost twin can linger in the survivor’s body for decades. One case documented chimerism from a vanished twin persisting 40 years later, identified when the surviving individual showed immune markers that could not have come from either parent.19PubMed Central. Cells from a vanished twin as a source of microchimerism 40 years later This kind of microchimerism adds a strange twist to the question of shared DNA: some people carry a tiny fraction of a twin’s genome inside them without ever knowing they had a twin at all.
What Spaceflight Revealed About Twin Divergence
NASA’s Twins Study offered a dramatic natural experiment. Astronaut Scott Kelly spent nearly a year aboard the International Space Station while his identical twin brother Mark stayed on Earth. Researchers tracked both men across a wide panel of biological measures. The spaceflight twin showed telomere elongation during flight, genome instability, shifts in DNA methylation in immune and oxidative stress pathways, altered gene expression, changes in gut microbiota, and some cognitive decline after returning to Earth.20PubMed Central. The NASA Twins Study: A multidimensional analysis of a year-long human spaceflight Most changes reversed within months of landing, but the study showed in real time how identical genomes can respond very differently when exposed to different environments. The underlying DNA sequence did not change in any large-scale way, but the regulatory landscape sitting on top of it shifted measurably.
Assisted Reproduction and Twinning Rates
Fertility treatments have changed the landscape of twinning. Transferring multiple embryos raises the chance of fraternal twins, but even transferring a single embryo carries a small risk of identical twinning if the embryo splits. Data from 2003 to 2012 showed monozygotic twinning rates of about 1.7 percent after day-2 or day-3 embryo transfers and about 2.5 percent after day-5 or day-6 transfers. Assisted hatching, a technique that thins the embryo’s outer shell, roughly doubled the risk of monozygotic twinning compared with transfers where it was not used.21PubMed Central. Trends and correlates of monozygotic twinning after single embryo transfer The natural rate of monozygotic twinning is roughly 0.4 percent, so the increase under assisted reproduction is notable, roughly four times higher for certain transfer protocols.22Human Reproduction. P-128 Morphokinetics and blastocyst biomarkers of twin and triplet monozygotic pregnancy using artificial intelligence
The genetics of fraternal twinning, meanwhile, are partly heritable. The tendency to release more than one egg per cycle runs in families through the maternal line. Large genome-wide studies have identified several genetic loci linked to dizygotic twinning, including regions near genes involved in follicle-stimulating hormone signaling and reproductive development.23PubMed Central. Genome-wide association study meta-analysis of dizygotic twinning illuminates genetic regulation of female fecundity Identical twinning, by contrast, does not appear to have a strong heritable component; it seems to be largely a random event during early embryo development.
Armadillo Quadruplets and the Limits of Genetic Identity
Humans are not the only mammals that produce genetically identical siblings. The nine-banded armadillo routinely gives birth to four genetically identical offspring from a single fertilized egg, making armadillo litters natural clones.24PubMed Central. What makes each of us unique? The nine-banded armadillo as a model to study individuality Researchers have used these quadruplets to study how individuality arises despite genetic identity. Blood RNA profiling across multiple litters revealed persistent gene-expression signatures unique to each individual within a set of clones, stable over time and even robust through experimental infection. In one litter, the most transcriptomically distinct individual showed an expanded population of inflammatory immune cells, a difference that persisted across multiple time points.25PubMed Central. Monocyte Lineage Expansion Drives Transcriptomic Individuality in Genetically Identical Armadillo Quadruplets
The armadillo work reinforces what the human twin data suggests: sharing a genome is not the same as being the same organism. Early stochastic events during development, essentially biological coin flips, produce lasting differences in cell composition and gene activity. The DNA code is the starting blueprint, but the way each body builds itself from that blueprint introduces individuality that no amount of genetic identity can erase.
Why the Assumption Matters for Twin Studies
Twin studies have been a cornerstone of genetics research for decades. The logic depends on comparing identical and fraternal twins: if identical twins are more similar for a trait than fraternal twins, the extra similarity is attributed to genetics. That logic assumes identical twins share 100 percent of their DNA and fraternal twins share 50 percent, treating both as clean reference points. It also assumes that both types of twins experience equally similar environments growing up. Simulations have shown that if genes even modestly influence the environments twins experience (because people who look identical get treated more alike), heritability can be overestimated by a substantial margin.26bioRxiv. Twin studies with unmet assumptions are biased towards genetic heritability Accounting for the genetic differences that actually exist between identical twins, and the environmental confounds that arise from their similarity, is an ongoing methodological challenge. It does not invalidate twin studies, but it means the headline heritability figures they produce are best understood as approximations rather than exact measurements.