Take two groups of mice. One group lives through a stressful stretch. The other doesn't. Then each group has pups, and the pups never meet their parents' stress at all. Yet the pups of the stressed mice come into the world anxious, with behaviour that looks like what their parents picked up.
Aleksey Shatskikh, a Candidate of Biological Sciences and senior researcher at the Laboratory of Functional Genomics at the Koltsov Institute of Developmental Biology, brought up this experiment on Scientific Russia when the host asked about something she had read: that a person's anxiety or depression can reach their children. How could a mood travel down a generation?
If that is possible, it changes how we think about what parents hand on. The genes are the same, but something else rides along with them. The mice answer one more question too. Fathers can send the signal, and fathers contribute only sperm.
The usual picture of epigenetics is a set of marks on the genes themselves. A chemical group sits on a gene, and the gene is silent. Take the group off, and the gene works. Other marks sit on the packaging around the DNA. Together they are a kind of markup laid over the text. That is the classic way these marks are thought of as being inherited.
Shatskikh's interest is in a different route. It has no marks on genes at all.
Cells make many regulatory RNAs that never become proteins. These non-coding RNAs act as guides. Each one finds its target by chemistry alone, drifting around the cell until it bumps into the right stretch of genome or the right RNA, and then it brings along a set of proteins. They switch genes on and off, and even steer the marking machinery to the right places. In the same lecture series he describes how one of them wraps a whole X chromosome in a "coat" and shuts it down.
Now suppose some of these RNAs end up inside a sperm cell or an egg. At fertilisation they enter the embryo along with everything else. Shatskikh's point is that the embryo gets a starting kit of regulators, and the mix and the amounts decide how it develops. Here is how he puts it:
Depending on which set of non-coding RNAs, and in what amounts, arrives with the sex cells into the new developing organism, this will determine how that embryo develops. И в зависимости от того, какой набор некодирующих РНК и в каких количествах поступит с половыми клетками в новый развивающийся организм, в зависимости от этого будет определяться то, каким образом этот зародыш будет развиваться.
— Aleksey Shatskikh, Scientific Russia, 17:41
What does that look like in the mice? Shatskikh says the nervous system develops differently. The pups' behaviour changes with it. The anxiety shows up after birth, and it resembles what the stressed parents showed. In the experiments he describes, small non-coding RNAs called microRNAs were found travelling this way, even from fathers through sperm, and they could set off the change in how the nervous system formed.
He is careful about the scope. Non-coding RNAs are not the only players in epigenetics, he says, and he makes no claim for people. What he describes is what was shown in mice.
Other labs have been probing the same channel. In October 2025, a paper in Nature Communications by Elizabeth Kleeman and colleagues injected RNA taken from the sperm of male mice infected with SARS-CoV-2 into fertilised eggs. The offspring showed a phenotype resembling that of pups born naturally to infected fathers. The authors read this as support for the idea that sperm RNAs contribute to the outcome. A June 2026 paper in Molecular Psychiatry by Pamudika Kiridena says that a father's infection before conception can change the offspring through his sperm's epigenetics, including small non-coding RNAs.
So the picture that emerges is odd. A father's experience can leave a trace in the few RNAs his sperm carries. Those RNAs enter an egg and meet a developing nervous system at the very start. The pup is then anxious from birth, without ever having had a reason to be.
Shatskikh calls this a special epigenetic mechanism, different from the classical marks. What is still open is how many such messages exist, and how far they reach.
