When it matters where you got your gene from: How 'parental origin' changes our traits

The same DNA letter can act differently depending on whether it came from your mom or dad. This is called the parent-of- origin effect (POE). A classic example is imprinting: in some parts of the genome, only the maternal or paternal copy of a gene works. A new study in Nature shows that such effects have significant effects on growth, metabolism, and other complex traits — and it did so in hundreds of thousands of people, even without their parents’ DNA.

Why is this important?

Most genetic studies assume a simple model: the effect of a variant depends only on how many copies of it you have (0, 1, or 2) — and it doesn’t matter who you inherited those copies from. But nature sometimes plays a more subtle game. According to the “parental conflict” hypothesis of evolution, paternal alleles are more likely to “push” offspring to grow taller and consume resources, while maternal alleles are more likely to conserve them. If this is true, we should see opposing effects of “mother’s” and “father’s” variants in traits related to growth and metabolism. Until now, there has been little convincing data on a broad range of traits: biobanks have the genotypes of participants, but usually not the genotypes of their moms and dads.

The Main Trick: How to Understand Where an Allele Came From Without the Parents' Genotypes

The authors proposed an elegant method of "surrogate parents". First, they stitch human chromosomes into two long haplotype "ribbons" - conventionally the "left" and "right" halves of the genome. Then they find out which of these ribbons most often coincides with a group of relatives on the maternal or paternal line. For this, they use:

• matches on the X chromosome in males and mitochondrial DNA (always maternal) to mark the "maternal side";
• information about sex differences in the recombination map of siblings to label regions as maternal or paternal;
• interchromosomal "phasing" across regions shared with first/second cousins in the biobank.

In this way, they were able to determine the parental origin of alleles for 109,385 UK Biobank participants – without a single parental genotype. They then verified the findings in the Estonian Biobank (up to 85,050 people) and the Norwegian MoBa cohort (42,346 children with their parents).

What exactly were you looking for?

The team conducted two large genome scans:

1. 59 complex traits (height, body mass index, type 2 diabetes, blood lipids, etc.) - comparing how much stronger each variant is if it is inherited from the mother versus the father.
2. >14,000 pQTL - genetic influences on blood protein levels.

The goal: to find areas where “mom’s” and “dad’s” copies produce different effects, even opposite ones.

Key Results

• More than 30 robust POE signals were found across traits and proteins, with a significant proportion in growth/IGF-1 and metabolism (e.g., type 2 diabetes and triglycerides). At more than a third of the loci, the effects of the “mother’s” and “father’s” alleles were in opposite directions, just as predicted by the conflict hypothesis.
• The validation performed impressively: ≈87% of the tested associations were confirmed in independent cohorts.
• The parent-less approach scales to biobanks: it increased the UK Biobank sample to ~109,000 people and, when combined with replications, yielded an analysis of up to 236,781 participants.

What does this mean in practice?

• Medical genetics. For a number of traits, predictions from polygenic models can be improved by taking into account who the allele is inherited from. Imagine two people with identical variants, but one got the “risk” from their mother, the other from their father. Their actual risks may differ, especially for metabolic phenotypes.
• Developmental biology. In real data in humans, we see the signature of a long-standing evolutionary "bargaining" between parental strategies: growth, energy, reserves. This is not just "textbook" imprinting; some POEs arise outside the classic imprinted clusters, which hints at additional mechanisms (regulation in trance, environmental influences, parental upbringing).
• Biobanks and epidemiology. Tools have emerged to learn POE in large datasets where familial genotypes are not available. This opens the way to re-evaluate known GWAS signals from the angle of maternal/paternal effects.

Important Disclaimers

• Although part of POE is explained by imprinting, not all of it - environmental channels (parental care, intrauterine factors) are also possible. It is difficult to separate them completely even with new methods.
• The effects, as in conventional GWAS, are small in size: they are strokes in a polygenic picture, not “fate switches.”
• The method requires good quality phasing and a sufficient number of relatives in the database; in populations where biobanks are smaller, “labeling” parents may be more difficult.

What's next?

Integrate POE into polygenic risks for specific diseases (type 2 diabetes, dyslipidemias) and test whether this improves risk stratification in the clinic. 2) Correlate POE loci with tissue-specific imprinting, methylation and expression maps to understand the mechanism. 3) Extend the approach to more diverse populations where relatedness patterns and allele frequencies are different.

Conclusion

This work shows convincingly that in human genetics, it’s not just the set of alleles that matters, but also where they come from. For a number of key traits, from height to lipid metabolism, parental origin really does change the equation. And now we have a massive way to account for this — even when the parents’ genotypes are nowhere to be found.