What Is a De Novo Genetic Mutation and What Causes It?

A de novo genetic mutation is a change in DNA that appears in a child but is not present in either biological parent’s own genome. The term comes from Latin, meaning “from new,” and these mutations arise spontaneously rather than being inherited in the traditional sense. Every person carries roughly 50 to 80 new single-letter DNA changes that neither parent had, and while most are harmless, de novo mutations are now recognized as a leading cause of severe early-onset developmental disorders. What generates them is not one thing but a convergence of biological processes, from simple copying errors during cell division to age-related changes in parental germ cells.

How DNA Copying Errors Create New Mutations

Your cells rely on molecular machinery to duplicate billions of DNA letters every time a cell divides. That machinery is impressively accurate, but not perfect. The fidelity of DNA replication depends on three error-catching steps working in sequence: the selectivity of the copying enzymes themselves, a built-in proofreading function that catches and removes mismatched letters, and a final mismatch-repair system that sweeps up errors the first two steps missed.¹ When all three work properly, the error rate is astonishingly low. But “astonishingly low” across three billion letters of DNA still means a handful of mistakes slip through with each generation.

Not all spots in the genome are equally vulnerable. Certain two-letter sequences called CpG sites are known mutational hotspots. The cytosine at these sites is often chemically modified by a methyl group, and methylated cytosine has a tendency to spontaneously lose an amino group and convert into thymine. This is a well-understood chemical reaction that happens constantly, and while repair enzymes usually catch it, the sheer frequency means some conversions escape correction.² A disproportionate share of disease-causing mutations in well-studied genes occur at these methylated CpG sites.³ So even without any environmental insult, the chemistry of your own DNA creates weak points where new mutations tend to cluster.

De novo mutations also concentrate in repetitive stretches of the genome. A recent four-generation pedigree study found that new single-letter mutations were enriched nearly threefold in centromeric regions and about twofold in segmental duplications compared to the genome average.⁴ These repetitive regions are harder for the copying machinery to navigate accurately, which makes intuitive sense if you think of a photocopier trying to reproduce a page full of the same sentence repeated over and over.

Why the Father’s Age Matters So Much

If there is one factor that dominates the rate of de novo mutations, it is the age of the father at the time of conception. Sperm-producing cells divide continuously throughout a man’s life, and each division is another chance for a copying error. A landmark Icelandic study that sequenced entire genomes of parents and children found that the number of new mutations increased by about two per year of paternal age, with an exponential model suggesting the count doubles roughly every 16.5 years.⁵ After accounting for normal random variation, the father’s age explained nearly all of the remaining differences in mutation counts between children.

A separate analysis of over 1,800 family trios confirmed this pattern, finding that new single-letter mutations accumulate with advancing paternal age at a rate of about 3% per year.⁶ To put that in perspective, a child conceived by a 40-year-old father will carry, on average, substantially more new mutations than a child conceived by a 20-year-old father. Most of these extra mutations land in harmless parts of the genome, but the sheer increase in numbers raises the odds that one lands somewhere consequential.

Selfish Sperm Cells and a Subtler Kind of Age Effect

The story of paternal age is not just about passive copying errors piling up. There is a more insidious process at work for a specific class of mutations. Certain mutations that occur in the stem cells responsible for making sperm actually give those stem cells a growth advantage, causing them to multiply faster than their neighbors. Over years, the mutant stem cells gradually crowd out normal ones, producing a larger and larger share of the sperm a man makes. Researchers have called this “selfish spermatogonial selection” because the mutation benefits the cell it arose in, even though it can cause serious disease in a resulting child.⁷

This process is thought to occur in the testes of all men and explains why certain rare but devastating conditions, such as Apert syndrome and achondroplasia, are almost always caused by mutations of paternal origin and become more common with older fathers.⁸ The mechanism is strikingly similar to how tumors grow: a single cell gains a growth advantage through a mutation and then expands clonally. In rare cases, this same process in the testes can even lead to testicular tumors.

Maternal Age Contributions

Mothers have traditionally been associated with chromosomal abnormalities like Down syndrome, which involve whole extra or missing chromosomes rather than single-letter mutations. The role of maternal age in point mutations was long considered negligible because egg cells do not keep dividing the way sperm-producing cells do. But more recent research has complicated that picture.

Studies have found that maternal age does influence the mutation rate, partly through the accumulation of DNA damage in egg cells that sit dormant for decades before being used, and potentially through effects on the number of mutations that arise in the early embryo after fertilization.⁹ An Icelandic study of over 1,500 family trios also showed that the types of mutations contributed by mothers shift with age in distinctive ways, with certain mutation signatures increasing and others decreasing per year.¹⁰ The maternal contribution is smaller in absolute numbers than the paternal one, but it is real and mechanistically distinct, driven more by accumulated chemical damage to DNA than by copying errors during cell division.

When in Development Do De Novo Mutations Arise

Not all de novo mutations happen in the egg or sperm before conception. Some occur after fertilization, during the rapid cell divisions of early embryonic development. These are called postzygotic or somatic de novo mutations, and they lead to a condition called mosaicism, where some cells in a person carry the mutation and others do not. De novo mutations can occur at essentially any life stage, from the parental germ cells through embryonic and fetal development, and even postnatally through aging.¹¹

One study that carefully reanalyzed 107 de novo mutations in 50 families found that about 6.5% of mutations initially assumed to be germline were actually present in a mosaic pattern in the child’s blood, indicating they arose after fertilization rather than in a parent’s egg or sperm.¹² Mosaicism matters clinically because a mosaic mutation may be milder than one present in every cell, and because standard genetic tests can miss it if the mutant cells are a small fraction of the sample.

Beyond Single-Letter Changes

When people hear “mutation,” they usually think of a single DNA letter swapped for another. But de novo mutations also include larger structural rearrangements: deletions, duplications, and inversions of whole chunks of DNA. These structural variants arise through different mechanisms than point mutations. One common route is called nonallelic homologous recombination, where the cell’s recombination machinery accidentally pairs up two similar-looking but non-matching stretches of DNA and swaps material between them, creating deletions on one copy and duplications on the other.¹³ Another route involves the cell’s machinery for repairing broken DNA strands, which can join the wrong ends together.¹⁴

These structural de novo mutations can be far more consequential than single-letter changes because they affect many genes at once. Well-known examples include the deletions and duplications that cause conditions like Williams syndrome, DiGeorge syndrome, and some forms of Charcot-Marie-Tooth disease. They tend to recur at the same genomic locations because the architecture of repetitive sequences in those regions makes them structurally fragile.

Do Environmental Exposures Cause More De Novo Mutations

Given that radiation and toxic chemicals can damage DNA, you might expect that a parent’s exposure to such agents would increase de novo mutations in their children. The evidence on this is more nuanced than most people assume. A small pilot study of offspring of soldiers exposed to high-frequency ionizing radiation found cases of unusually high rates of a specific type of clustered mutation, suggesting radiation exposure can leave a detectable mark on the germline.¹⁵

Chemotherapy, however, tells a more surprising story. A study of testicular cancer survivors found no increase in de novo germline mutations in children conceived after treatment compared to siblings conceived before treatment.¹⁶ Research on specific “selfish” mutations in sperm after chemotherapy found that highly sterilizing treatments actually reduced the levels of certain age-related mutations, probably because the treatment wiped out the expanded clones of mutant sperm stem cells and forced the testes to regenerate from a smaller pool of surviving cells.¹⁷ This counterintuitive finding suggests that not all DNA-damaging exposures translate into heritable mutations in the same way, and that the biology of sperm stem cell recovery after injury is more complicated than a simple “more damage equals more mutations” model.

Connections to Autism, Schizophrenia, and Developmental Disorders

De novo mutations have emerged as one of the most important genetic explanations for conditions that seem to appear out of nowhere in families with no history of the disorder. Severe intellectual disability, autism spectrum disorder, and other developmental conditions are now understood to frequently arise from new mutations rather than inherited ones.¹⁸ This makes evolutionary sense: conditions that severely reduce the likelihood of reproduction cannot be maintained in a population through inheritance alone. They persist because they keep being generated fresh each generation.

In autism, de novo mutations have been recognized as a strong source of genetic causality, with research mapping which genes and pathways these mutations tend to hit.¹⁹ In schizophrenia, exome sequencing of families with no prior history of the condition found a roughly 3.5-fold increase in the proportion of protein-truncating de novo mutations compared to what would be expected by chance, with several of the affected genes overlapping those already implicated in autism and intellectual disability.²⁰ The convergence on shared genes across these different diagnoses suggests that some fundamental brain-development processes are especially sensitive to disruption by new mutations.

Mutations Outside of Genes

Most clinical genetic testing focuses on the roughly 1.5% of the genome that codes for proteins. But de novo mutations also land in the vast noncoding territory that regulates when, where, and how much those genes are turned on. A large study of neurodevelopmental disorders found a significant enrichment of de novo mutations in highly conserved regulatory elements that are active in the fetal brain, with an estimated 1 to 3% of patients who lack a diagnostic coding mutation carrying a pathogenic de novo mutation in one of these regulatory regions instead.²¹

That percentage may sound small, but for families who have gone through extensive genetic testing without finding an answer, it represents real diagnostic territory that has only recently become accessible. Research in autism has shown that noncoding de novo mutations near genes co-expressed in the developing brain with known autism risk genes are more likely to fall in functionally important regions in affected children than in their unaffected siblings.²² The challenge is that interpreting these noncoding mutations is far harder than interpreting a mutation that breaks a protein. A regulatory mutation might subtly change the volume knob on a gene rather than destroying the gene outright, making it difficult to prove that any single noncoding change is the culprit.

How De Novo Mutations Are Found

Detecting a de novo mutation requires sequencing the DNA of a child and both parents, then looking for changes present in the child but absent in both parents. This trio-based approach sounds straightforward, but the technical challenge is significant. False positives are common because sequencing errors can mimic real mutations, and mosaic mutations in a parent can be missed if they are present in only a small fraction of cells. Modern computational tools have achieved precision rates above 98% after careful filtering.²³ Different software tools produce substantially different numbers of candidate mutations from the same data, which is why most clinical and research settings now use multiple tools and take the overlap as a higher-confidence set.²⁴

For families seeking a diagnosis, exome or genome sequencing of the trio is increasingly offered as a clinical test. When a clearly damaging de novo mutation is found in a gene known to cause disease, it can provide a definitive answer. But the process also generates uncertainty: variants of unknown significance, mutations in genes not yet linked to any condition, and the challenge of interpreting results in the context of prenatal decision-making all create real counseling difficulties.²⁵

The Evolutionary Balancing Act

De novo mutations are the raw material of evolution. Without them, there would be no new genetic variation for natural selection to act on. The human germline mutation rate sits at roughly 1.2 × 10⁻⁸ new single-letter changes per DNA position per generation, which translates to those 50-80 mutations per child mentioned earlier.⁴ Natural selection works to remove the harmful ones: an analysis of over 70,000 human genomes estimated that about 0.4 to 0.7% of the genome is under extreme purifying selection, implying that each generation produces roughly a quarter to half a strongly harmful new mutation per person on average.²⁶

Comparing humans to other primates offers a useful window into how these processes scale. A study comparing humans and baboons found that baboons have a per-generation mutation rate roughly half that of humans, yet both species show a similar fourfold male bias in mutation origin.²⁷ The consistency of that male bias across species with very different reproductive timelines challenges the simple assumption that mutations track cell divisions in a linear way. Something about the male germline beyond sheer number of divisions appears to make it inherently more mutagenic, though exactly what that is remains an active area of research.