crosscalc

Did a 2026 mouse study break Mendel's laws? Explained

In the news Published

Table of contents
  1. First, what Mendel’s laws predict
  2. What is DNA methylation?
  3. What the researchers did
  4. What they found
  5. What the study did not show
  6. So, is Mendel wrong?
  7. Common questions
  8. Did scientists prove Mendel wrong?
  9. Does epigenetic mean not genetic?
  10. Does this study apply to my family?
  11. Check your understanding
  12. Related
  13. Sources

Level: High school (advanced) to college · You’ll need: genotypes and alleles and a monohybrid cross

In May 2026, researchers at Johns Hopkins and Texas A&M reported that about 7% of the inherited DNA methylation patterns they tracked on the non-sex chromosomes (autosomes) in mice did not follow Mendel’s laws. Headlines said scientists had found “inherited traits that break Mendel’s laws”. That’s partly true and partly overstated. The DNA itself was still inherited the Mendelian way, about 93% of methylation patterns followed Mendel too, and the study measured chemical marks on DNA, not visible traits like coat color.

This page explains what the study did, what “non-Mendelian” means here, and how it fits with the Punnett squares you already know.

First, what Mendel’s laws predict

Mendel’s laws describe how alleles (versions of a gene) pass from parents to offspring:

  • Segregation: each parent has two copies of a gene and passes on one, chosen at random.
  • Independent assortment: genes on different chromosomes are sorted into gametes independently of each other.

That’s what a Punnett square models. In Aa × Aa, each parent passes on A or a with a ½ chance, so ¼ of offspring are aa. You can check this in the monohybrid calculator.

The key idea for this study: in a Mendelian pattern, a feature travels with the allele that carries it. If you know which copy of a chromosome an offspring inherited, you can predict the feature.

What is DNA methylation?

DNA methylation is a small chemical tag (a methyl group) attached to DNA, usually at C–G sites. It doesn’t change the DNA sequence (the A, C, G and T letters). Instead, it’s one of the ways cells control which genes are active. This kind of control, layered on top of the sequence, is called epigenetics.

Methylation patterns are often described as epigenetic marks. Most are reset between generations, which is one reason the question “are they inherited the same way as genes?” is interesting.

What the researchers did

The team worked with two inbred mouse strains, CC019 and CC037, from a research panel called the Collaborative Cross. Inbred strains are useful because every mouse in a strain is (almost) genetically identical, a bit like a pure-breeding pea line in Mendel’s experiments.

  1. Generation 1 (P): the two parent strains.
  2. Generation 2 (F1): offspring from crossing the strains both ways (CC019 mother × CC037 father, and the reverse).
  3. Generation 3 (F2): offspring from crossing F1 mice with each other.

They used long-read nanopore sequencing, which reads the DNA sequence and its methylation on the same stretch of DNA. That let them tell which methylation belonged to the copy inherited from the mother and which to the copy from the father. This is called allele-specific methylation. They studied liver and muscle tissue (the F2 analysis used liver) and treated the methylation level at each region like a trait to follow through the generations.

What they found

  • About 93% of autosomal methylation patterns were Mendelian. The methylation followed the local DNA, so you could predict it from which allele was inherited. In most cases, a nearby DNA difference was responsible.
  • About 7% were not, at least 522 regions. These sorted into several types (the paper defines 12 patterns in all).
  • 54 “emergent” patterns. In these, F1 mice showed methylation on at least one allele that neither parent strain had.
  • A few new imprinted genes. In genomic imprinting, whether a gene copy is active depends on which parent it came from. The team reported five seemingly new imprinted genes.
  • Paramutation in a non-engineered mammal. In paramutation, one allele changes the state of the other allele in the same individual, and the change can then be passed on. It was first described in maize. In mice, it had only been reported before in engineered (transgenic) mice. Here it was confirmed in the gene Capn11 and was highly likely in Vps37c, in both cases linked to DNA sequences called transposable elements.

What the study did not show

This is where some headlines went too far.

  • No visible traits were measured. The “traits” were methylation levels (and some gene activity in the liver). The study didn’t show a coat color, size or disease that skips Mendel’s rules.
  • The DNA sequence was still Mendelian. The non-Mendelian part is how methylation was passed on and set up, not how alleles were.
  • “Non-Mendelian” is a broad label here. Some of the 522 cases were caused by DNA differences elsewhere in the genome affecting methylation at a distance. That’s still genetics, just not the local allele.
  • It was mice, not people. The authors note there’s no comparable full analysis in humans yet, and human studies are harder because of differences in environment, age and genetic background.
  • It’s a first look. Only two strains and two tissues were studied, and about a third of the genome couldn’t be assigned to the mother’s or father’s copy. The authors suggest combining this approach with physical measurements in future work.

So, is Mendel wrong?

No. Mendel’s laws still describe how alleles pass from parents to offspring, and this study found the large majority of methylation patterns follow them too. What it adds is evidence that a meaningful minority of epigenetic marks are inherited or set up in more complicated ways.

A good way to think about it: a Punnett square is a model. It works when its assumptions hold (one gene, complete dominance, alleles passed on unchanged). Studies like this one map where those assumptions stop. For more examples, from blood types to height, see beyond Mendelian inheritance.

Common questions

Did scientists prove Mendel wrong?

No. They found exceptions in methylation, not in how DNA is inherited.

Does epigenetic mean not genetic?

No. Many methylation patterns here were controlled by DNA differences, either nearby or far away.

Does this study apply to my family?

No. The study was in mice, in two tissues, and measured no visible traits.

Check your understanding

  1. In this study, what does “Mendelian” mean for a methylation pattern?
  2. Roughly what share of autosomal methylation patterns did not follow Mendel?
  3. What makes the paramutation finding new?
  4. Name one thing the study did not measure.
Answers
  1. The methylation follows the local DNA allele, so knowing which chromosome copy an offspring inherited predicts its methylation.
  2. About 7% (at least 522 regions). About 93% were Mendelian.
  3. Paramutation had been reported in mice only in engineered (transgenic) mice. Here it appeared in ordinary strains, without genetic engineering.
  4. Visible or physical traits (such as coat color or size). It also didn’t study humans.

Sources

  • Davidovich A, Cuomo D, Su H, et al., Hansen KD, Threadgill DW, Feinberg AP. “Non-Mendelian inheritance of DNA methylation patterns in mice.” Nature Genetics 58, 1409–1422 (2026). doi:10.1038/s41588-026-02604-z · open full text (PMC)
  • Johns Hopkins Medicine news release via ScienceDaily, “Scientists discover inherited traits that break Mendel’s Laws of genetics” (posted May 30–June 1, 2026). sciencedaily.com
  • Chandler VL. “Paramutation: from maize to mice.” Cell 128, 641–645 (2007). doi:10.1016/j.cell.2007.02.007
  • MedlinePlus Genetics (U.S. National Library of Medicine), “What is the epigenome?” medlineplus.gov
Try it: work out any cross with the mouse coat color calculator, or a general one with the Punnett square calculator.