Methylation tags in offspring that neither parent carried
For over a century, Gregor Mendel's arithmetic of inheritance has served as biology's most reliable compass — yet a team at Johns Hopkins University has now charted territory it cannot reach. By tracking chemical tags on DNA across three generations of laboratory mice, researchers discovered that some offspring carry molecular markers absent in both parents, and that one gene's tag can silently copy itself onto its paired counterpart without any alteration to the underlying code. These findings suggest that the story of heredity has a quieter, more responsive chapter than classical genetics ever imagined — one written not in the sequence of life, but in the annotations layered upon it.
- Seven percent of methylation sites across the mouse genome broke Mendel's rules entirely, with 54 sites showing chemical tags in offspring that neither parent possessed.
- Paramutation — a phenomenon where a tag on one gene copy spreads to the other — appeared naturally in unmodified mice for the first time, upending a pattern previously seen only in plants, flies, and deliberately engineered animals.
- One of the affected genes governs sperm development in both mice and humans, pulling the discovery into the contested and consequential terrain of fertility research.
- Because methylation can be shaped by diet, stress, and environmental exposure, these non-Mendelian patterns could allow traits to propagate through populations far faster than waiting for a random mutation in the DNA sequence itself.
- The research team is now turning the same long-read sequencing technique toward human family data, and clinicians are being urged to look beyond standard genetic tests when inherited conditions resist obvious explanation.
For more than a century, inheritance has been taught as a clean arithmetic — one gene from each parent, combined in predictable ratios, following the rules Gregor Mendel discovered in his pea plants. Two parents without a chemical tag on a stretch of DNA should produce offspring without that tag. Except, as a team led by Dr. Andrew Feinberg at Johns Hopkins recently found, that is not always what happens.
Feinberg and colleagues bred three generations of laboratory mice and used long-read sequencing to track DNA methylation — chemical tags that attach to DNA without altering the genetic code, functioning as switches that turn genes on or off. The technique captured both the genetic sequence and its methylation pattern on the same molecules, revealing far longer stretches of DNA than older methods allowed.
About 93 percent of methylation patterns followed Mendel's rules in some form. But 522 sites did not, and 54 of those showed tags in offspring that were absent in both parents entirely. The phenomenon responsible is called paramutation: methylation on one copy of a gene somehow copies itself onto the paired copy, so both are tagged even when only one parent contributed the marker. This had been documented before — in plants, flies, and genetically engineered mice — but never naturally, in animals with no deliberate genetic alteration. One of the genes where it occurred is involved in sperm development in both mice and humans, a detail that places the finding within reach of fertility research.
The team also identified five additional genes governed by genomic imprinting, a known exception to Mendel's rules where a tag from one parent silences that parent's copy of a gene — expanding what had been a sparse catalog in the mouse genome.
The broader implication is striking: because methylation can be shaped by diet, stress, and environmental exposure, epigenetic changes could spread through a population faster than waiting for a random mutation in the DNA sequence itself. Feinberg's team now plans to search for the same patterns in human family data. For clinicians, the practical message is direct — when a condition runs through a family without a clear genetic cause, the answer may live in a layer that standard sequencing tools never read. The study appears in Nature Genetics.
For more than a century, we have taught inheritance as a clean arithmetic: one gene from each parent, combined in predictable ratios, producing offspring that follow the rules Gregor Mendel discovered in his pea plants. Two mice without a chemical tag on a particular stretch of DNA should produce offspring without that tag. The math is simple. The outcome is reliable. Except, as a team of researchers recently discovered, it often is not.
Dr. Andrew Feinberg and colleagues at Johns Hopkins University School of Medicine, working with researchers at Texas A&M, bred three generations of laboratory mice—26 in the founding generation, 34 in the second, and 19 in the third—to track something Mendel could never have seen: DNA methylation. These chemical tags attach to DNA without altering the genetic code itself, functioning as switches that turn genes on or off. They sit on either copy of a gene and pass from parent to child. The researchers sampled liver and muscle tissue from each animal and used long-read sequencing to capture both the genetic sequence and the methylation pattern on the same DNA molecules, a technique that reveals far longer stretches of DNA than older methods allowed.
The results were mostly unsurprising. About 93 percent of the methylation patterns followed Mendel's classical rules in some form. But the remaining 7 percent—roughly 522 sites across the genome—did not. More strikingly, 54 of those sites showed methylation tags in offspring that were nowhere to be found in either parent. A pairing between two mice with no methylation on a particular spot should produce offspring with no methylation on that spot. Instead, the team kept finding offspring with methylation on both copies of the gene.
The phenomenon at work is called paramutation, and it is strange enough that it has been documented before—but only in plants, flies, and genetically engineered mice. In those engineered cases, scientists had deliberately altered the genome to produce the effect. This was the first time paramutation had appeared naturally in mice with no genetic tampering. The methylation on one copy of a gene somehow gets copied onto the other copy, so both are tagged even when only one parent contributed the tag. One of the genes where this occurred helps drive normal sperm development in mice and humans. Earlier research has linked reduced function of the human version to infertility, placing this finding in territory worth tracking for fertility research, though not as a definitive link.
Feinberg's team also identified five additional genes that behave according to genomic imprinting, a well-known exception to Mendel's rules where a tag from one parent silences that parent's copy of a gene. The silencing depends on which parent contributed the copy, not on which version is stronger or weaker. This expands what had been a thin catalog of such genes in the mouse genome.
The implications reach beyond the laboratory. Methylation tags can come from outside the genome entirely. Diet, stress, and environmental exposure have all been linked to changes in methylation patterns in earlier research, and some of those changes can pass down to offspring. This makes the new findings unusual in a particular way: epigenetic inheritance that changes in response to the environment could spread through a population faster than waiting for a random change in the DNA sequence itself. As Feinberg noted, non-Mendelian patterns of inheriting epigenetics could be a faster way to acquire diverse or new traits than alterations in the genomic sequence, especially in response to environmental pressures.
The team now plans to apply the same approach to human genomic data, looking for similar patterns in families to confirm whether this kind of inheritance occurs in people the same way it was just found in mice. For clinical geneticists, the message is practical: when standard genetic explanations fall short, the answer may sit in a layer that standard sequencing tools never read. If a condition runs through a family without an obvious genetic cause, the culprit may be methylation. The study appears in Nature Genetics.
Citações Notáveis
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— Dr. Andrew Feinberg, Johns Hopkins University School of Medicine
A Conversa do Hearth Outra perspectiva sobre a história
So if 93 percent of methylation patterns follow Mendel's rules, why does the 7 percent matter so much?
Because it's the 7 percent that tells us inheritance is more complicated than we thought. Those 522 sites represent places where the textbook breaks down. And some of those sites show methylation in offspring that neither parent had—that shouldn't be possible under classical genetics.
How does a tag appear if neither parent carried it?
That's paramutation. The methylation on one copy of a gene somehow spreads to the other copy. Both copies end up tagged even though only one parent contributed a tag. It's like the mark is contagious within the cell.
Has this been seen before?
In plants, flies, and engineered mice, yes. But never in unmodified mice living naturally. That's the breakthrough. It means this isn't just a laboratory artifact—it's something that happens in real genomes without human intervention.
What does this have to do with the environment?
Methylation tags respond to diet, stress, and exposure. If those environmental changes can be inherited and spread through paramutation, then organisms could adapt to new conditions much faster than waiting for random DNA mutations. It's a quicker evolutionary tool.
So a mouse stressed by its environment could pass that stress down in a way that changes how its offspring's genes work?
Exactly. And if that epigenetic change spreads through paramutation, it could reshape a population's traits in a generation or two, not over many generations of genetic drift.