For more than a century, heredity has been framed through the tidy logic of Mendel’s pea plants: traits pass from parent to offspring by fixed genetic rules. But a new mouse study suggests that chemical marks layered on DNA can sometimes slip past those rules, carrying inherited effects in ways standard genetics does not fully explain.

That does not mean the basic laws of genetics are suddenly obsolete. It does mean they may not tell the whole story. In this case, researchers found that DNA methylation, a chemical tag that can turn genes on or off without changing the DNA sequence itself, sometimes passed across generations in ways that did not follow the classic patterns of dominance, recessiveness, or simple parental contribution.

Looking across three generations of mice, the team identified 522 cases on non-sex chromosomes where inherited methylation patterns broke Mendel’s rules. Those cases made up about 7% of the epigenetic inheritance patterns they tracked, a share large enough to suggest these exceptions are not just biological curiosities.

“Non-Mendelian patterns of inheriting epigenetics could be a faster way to acquire diverse or new traits than alterations in the genomic sequence itself, especially in response to environmental pressures,” said Andrew Feinberg, a Bloomberg Distinguished Professor at Johns Hopkins University and co-leader of the research.

The study, published in Nature Genetics, adds weight to a long-simmering idea in biology: that inheritance is shaped not only by DNA sequence, but also by chemical markings that can influence how genes behave.

Mendel’s laws describe how gene variants, known as alleles, are passed from parents to offspring. One allele comes from each parent, and the interaction between those variants helps determine which traits appear. The framework has guided genetics for generations because it works well for a large share of inherited traits.

But epigenetics adds another layer. DNA methylation does not rewrite the genetic code, yet it can influence whether a gene is active or silent. Some forms of epigenetic inheritance have already been shown to bend standard expectations. Genomic imprinting, for example, can silence an allele based on whether it came from the mother or father, rather than whether it is dominant or recessive.

The new study found familiar examples of imprinting, but also several other inheritance patterns that were harder to fit into the usual categories. Some appeared to be controlled by nearby genetic variation, while others seemed to be shaped by distant regulatory influences or by interactions between paired alleles themselves.

To sort that out, the researchers studied two genetically distinct inbred mouse strains from the Collaborative Cross, a model designed to capture a wide range of genetic diversity. They examined methylation in liver and muscle tissue from 26 mice in the parental generation, 34 in the F1 generation, and 19 in the F2 generation, all between 4 and 6 months old.

Using long-read Oxford Nanopore sequencing, the team measured both DNA sequence and methylation on the same long DNA molecules. That mattered because it let them tell which methylation marks belonged to which allele, a level of detail that short-read methods often miss.

Most of the inheritance patterns the team found were still consistent with ordinary genetics. In fact, the largest category involved 7,081 genomic regions in which methylation tracked with nearby genetic variation, known as cis-acting methylation quantitative trait loci, or meQTLs.

But the most intriguing results came from the exceptions.

The researchers found 54 examples of what they called emergent epigenetic inheritance patterns, cases in which the offspring showed a methylation pattern not seen in either parent. In one striking scenario, two mice lacking methylation on the same allele could produce offspring with methylation on both alleles.

“The methylation seemingly appeared out of nowhere,” Feinberg said.

That kind of result hints at a mechanism for generating biological diversity without first changing the DNA sequence itself. The authors say such patterns could help explain puzzling features of inheritance, including incomplete penetrance, unusual family patterns of disease, and traits shaped by environmental exposure.

They also found at least 51 regions under the control of distal trans-acting meQTLs, where the factors influencing methylation appear to lie far from the affected DNA region. Other regions showed dominant or mixed methylation states that did not line up with standard Mendelian expectations.

One of the study’s most unusual findings was evidence for a naturally occurring paramutation in a mammal.

Paramutation is a form of inheritance in which the methylation state of one allele appears to alter the methylation state of its partner. It has been reported in plants and flies, and in engineered or transgenic mice, but not as a naturally occurring example in a mammalian genome.

The clearest case involved a gene called Capn11, which helps regulate normal sperm development. In humans, altered expression of the related gene has been linked to infertility and sperm abnormalities.

“It’s almost like the methylation is transferred to another allele,” Feinberg said.

The team also found two additional highly likely examples of paramutation associated with IAP elements, a type of endogenous retroviral sequence in the genome. These repetitive elements are already known to be sensitive to epigenetic control and, in some cases, to environmental influence. Their presence in these regions suggests they may help drive some of the non-Mendelian inheritance patterns the researchers observed.

That matters because it points to a system that is not only more flexible than classical genetics, but potentially more responsive to outside pressures as well. Feinberg noted that epigenetic influences on the genome have been linked to environmental stress, trauma, and diet.

The study does not overturn Mendel. Instead, it expands the biological picture around him.

For many inherited traits, standard genetics still works. But the new results suggest that if researchers focus only on DNA sequence, they may miss an important part of how traits, disease risks, and gene activity move through families.

Kasper Hansen, a professor of biostatistics at Johns Hopkins and a co-corresponding author, said the findings could push the field toward a more integrated view. “This work may convince scientists to integrate both genomics and epigenomics more often for a complete understanding of how traits that produce disease and healthy states are inherited,” he said.

The team also tied some methylation patterns to gene expression. They identified 700 genes showing both cis-acting regulation of DNA methylation and gene expression, along with smaller sets linked to trans-acting, dominant, emergent, imprinting, and sex-specific methylation patterns. That does not prove every inherited methylation mark changes a trait, but it strengthens the case that at least some of them have real biological effects.

The work also has limits. It examined only two mouse strains, only two tissues, and only animals raised under stable conditions. The authors say broader studies, involving more tissues, environments, and eventually human data, will likely reveal even more complicated inheritance patterns.

The findings give geneticists another reason to look beyond DNA sequence alone when family patterns do not make sense.

If similar inheritance patterns exist in humans, they could help explain why some diseases or traits appear in ways that standard gene testing does not predict.

The work also points toward a future in which researchers combine genomic and epigenomic analysis more routinely, especially when studying disease risk, environmental effects, and inherited conditions that do not follow clean Mendelian lines.