A recent experiment with cloning mice and the subsequent fertility problems that resulted are directly related to the fertility problems predicted by THE FROZEN GENE and which serve as the basis for the BIOSTELLAR novels. To be honest, it’s been a bit disappointing that people haven’t paid anywhere nearly as much attention to the core problem addressed by TFG as they have to the disproof of evolution by natural selection, especially now we know that the problem I identified in the human genome on the basis of the math and logic has a sound basis in empirical scientific data as well.
A twenty-year experiment at the University of Yamanashi just ended in the most predictable way imaginable—if you’ve been paying attention to the math. A team led by developmental biologist Teruhiko Wakayama took a single female mouse in 2005, cloned her, then cloned the clone, then cloned the clone of the clone, and kept going for two decades. Over 30,000 cloning attempts. More than 1,200 mice produced. Fifty-eight generations from that single original donor.
If you’ve read SIGMA GAME, then you probably already know what was coming: model collapse.
Generation 58 was the last. Every mouse born from it died within days.
The study was published in Nature Communications on March 24, 2026 (Wakayama et al., Nat. Commun. 17, 2495), and the results are worth walking through carefully, because they have implications that reach considerably further than the cloning industry.
For the first 25 generations, everything looked fine. The cloned mice were healthy, had normal lifespans—about two years, which is standard for a lab mouse—and the success rate of the cloning procedure was actually improving. At one point, the researchers themselves speculated that serial cloning might be sustainable indefinitely.
Then the success rate started dropping. By the 57th generation, the birth rate had fallen to a fraction of a percent. The genome sequencing told the story: approximately 3,700 single-nucleotide variants and 80 insertion-deletion mutations had accumulated across the lineage, averaging about 69 new mutations per generation—roughly three times the rate you see in sexually-reproduced mice. The frequency of deleterious mutations had nearly doubled. Some animals had lost an entire X chromosome. Others showed translocations—pieces of chromosomes breaking off and reattaching to the wrong partner.
By the 58th generation, the accumulated damage was lethal. The mice looked physically normal at birth but were too broken at the genomic level to survive.
Wakayama’s own summary was characteristically understated for a Japanese scientist: “We had believed that we could create an infinite number of clones. That is why these results are so disappointing.”
But if they’d read THE FROZEN GENE, then they wouldn’t have been surprised.
What the Yamanashi team documented over twenty years is a textbook demonstration of a concept called Muller’s Ratchet, proposed by the geneticist Hermann Muller in 1964. The idea is simple: in any lineage that reproduces without sexual recombination, deleterious mutations can only accumulate. They never get removed. The ratchet turns in one direction—toward deterioration—and it never turns back.
In normal sexual reproduction, two parents contribute DNA, and the resulting offspring gets a shuffled combination of both. This shuffling does two things: it occasionally concentrates deleterious mutations into a single unlucky offspring who fails to reproduce (purging the bad variants from the population), and it occasionally produces offspring with fewer deleterious mutations than either parent. The net effect is that sexual reproduction acts as a quality-control mechanism, keeping the genome roughly stable over time.
Remove the shuffling, and you remove the quality control. Every copying error persists. Every mutation that doesn’t immediately kill the organism gets passed to the next generation, along with whatever new mutations that generation accumulates. The genome doesn’t improve. It doesn’t even hold steady. It deteriorates, generation by generation, until the accumulated damage exceeds whatever threshold the organism requires for viability.
That’s exactly what happened to the Yamanashi mice. The ratchet clicked 58 times, and then the line was dead.
This is not a new prediction. It’s not a controversial hypothesis. It is a straightforward mathematical consequence of copying without error correction, and it has been understood theoretically for sixty years. What’s new is that someone finally ran the experiment in mammals for long enough to watch it happen in real time.
The most interesting result in the Wakayama study isn’t the collapse. It’s what happened when they tried to rescue the dying lineage.
The researchers took female mice from the 50th and 55th generations—deep into the damage zone—and mated them with normal, sexually-reproduced male mice. The first generation of offspring was still somewhat compromised: smaller litters, the characteristic oversized placentas that plague cloned animals, and some of the accumulated genetic problems. But the grandchildren—just two generations of sexual reproduction later—were completely normal.
Two generations. Twenty years of accumulated genetic damage, reversed in two generations of sex.
That sounds like great news for sexual reproduction, and it is. But think carefully about why it worked, because this is where the result connects to something larger.
The reset worked because one side of the pairing was genetically healthy. The normal males had come from a population maintained by continuous purifying selection—a population where individuals carrying heavy mutational loads had been dying or failing to reproduce for their entire evolutionary history, keeping the genome clean. The sexual recombination between a damaged clone and a healthy male allowed the offspring to inherit clean copies of the genes that had been corrupted in the clonal lineage. The damaged variants were either not inherited or were masked by functional copies from the father’s side. By the second generation, the worst of the damage had been shuffled out.
The key phrase in that explanation is one side of the pairing was genetically healthy. Here is the key phrase: The reset required a clean genome to reset against.
For those of you who have read Probability Zero, you know that my co-author Claude Athos and I also published a series of companion papers under the title The Frozen Gene. Where Probability Zero demonstrates the mathematical impossibility of the Theory of Evolution by Natural Selectio, The Frozen Gene addresses a different but related question: what is actually happening to the human genome right now?
The answer, supported by both theoretical derivation and empirical analysis of ancient DNA, is: nothing. The human genome is frozen.
The core concept is something we call the Selective Turnover Coefficient, denoted d. It measures the fraction of a population that is replaced by selection each generation—essentially, how much “room” exists for natural selection to operate. During the Paleolithic, when human life was nasty, brutish, and short, d was approximately 0.66. During the Neolithic, after agriculture but before modern medicine, it was around 0.53. A beneficial allele that would have reached fixation in about 18,000 years under Neolithic conditions now requires approximately 630,000 years under contemporary demography.
Today, d is approximately 0.015.
That is not a typo. The selective turnover of the human population has collapsed by a factor of 35 from its Neolithic baseline, and by a factor of 44 from its Paleolithic baseline. Nearly everyone born today survives to reproductive age. Nearly everyone who reproduces successfully raises offspring who survive to reproductive age themselves. The differential survival and reproduction that natural selection requires in order to operate has been almost entirely eliminated by modern medicine, modern sanitation, and modern food systems.
This means the gene pool is effectively frozen in place. Beneficial alleles cannot spread because there is no differential reproduction to spread them. The same mechanism that prevents beneficial change also prevents the efficient purging of deleterious mutations. The sieve that kept the genome clean for hundreds of thousands of years has been switched off.
Now consider the Yamanashi cloning experiment in light of The Frozen Gene.
The cloning experiment represents the extreme case: d = 0, or as close to it as physically possible. No recombination. No selective mixing. No error correction of any kind. Pure copying, generation after generation, with every error preserved and compounded. Result: genomic collapse in 58 generations.
Modern humanity represents a less extreme but structurally identical situation. We still have sexual recombination—the shuffling mechanism still operates—but the selective component that makes recombination effective at purging damage has largely been disabled. Recombination can concentrate deleterious mutations into a single individual, but if that individual survives and reproduces anyway because modern medicine keeps them alive and modern society ensures their offspring survive, the purging doesn’t happen. The mutations stay in the population.
Remember the two-generation reset from the Wakayama experiment. It worked because the healthy males came from a population where selection was still operating—where loaded individuals were still being removed. That’s the critical ingredient, and it’s the ingredient that’s disappearing.
When d was 0.53, the reset mechanism worked. Individuals with heavy mutational loads died young or failed to reproduce, and their variants were purged from the population. When d is 0.015, the reset mechanism is still technically present, but it’s operating at roughly 3% of its historical power. The sieve has holes in it so large that almost everything passes through.
We confirmed this empirically. Using the Allen Ancient DNA Resource—17,629 ancient European genomes spanning 8,000 years—we measured the ratio of derived allele burden at constrained sites (sites where mutations are likely to be harmful) versus neutral sites (sites where mutations don’t matter) across the entire Holocene. The ratio was flat from the Early Neolithic through the Bronze Age, when d was stable. It began rising during the Iron Age, when d started declining. And it jumped sharply in modern populations, where d has collapsed.
The constrained-site burden in modern Europeans is 3.47% higher than in medieval Europeans, with a statistical significance of p = 10⁻⁶². Neutral sites didn’t move. Only the constrained sites—the ones that purifying selection is supposed to keep clean—showed the increase. That’s not noise. That’s relaxed purifying selection, measured directly in the ancient DNA record.
The Yamanashi experiment gives us three things we didn’t have before.
First, it provides a direct experimental confirmation that Muller’s Ratchet operates in mammals under controlled laboratory conditions. This was theoretically predicted and had been inferred from natural populations—the last woolly mammoths on Wrangel Island almost certainly died from exactly this kind of mutational meltdown—but it had never been demonstrated experimentally in mammals across a sufficient number of generations to observe the complete trajectory from health to extinction. Now it has.
Second, it gives us a concrete number: 58 generations from a healthy starting genome to total lineage failure under zero recombination. That’s not directly applicable to the human situation, where recombination still occurs, but it establishes the timescale of the extreme case. When the sieve is completely off, mammalian genomes last less than 60 generations. The question for modern humanity is not whether the sieve being mostly off will produce the same result. It’s how much longer the trajectory takes when you retain recombination but disable selection.
Third, and most importantly, the two-generation reset confirms that the power of sexual reproduction to correct genetic damage depends entirely on the selective maintenance of the population it draws from. Sex is not magic. Recombination is not, by itself, a genome-cleaning mechanism. It is a shuffling mechanism, and shuffling only cleans the genome if selection subsequently removes the individuals who drew bad hands. When selection stops removing those individuals, the shuffling just redistributes the damage without reducing it.
That is the situation modern humanity is entering. The shuffle still works. The discard pile is closed. But there are still some uncomfortable implications.
Every generation, each human being is born with approximately 70 new mutations, of which roughly 2.2 are meaningfully deleterious. For the entirety of human history prior to about 1900, those deleterious mutations were balanced by purifying selection—loaded individuals dying before reproduction, or reproducing less successfully. The genome stayed roughly stable because the input of new damage was matched by the selective removal of old damage.
Since 1900, the input has continued at the same rate. The removal has effectively stopped. We are accumulating approximately 2.2 new deleterious mutations per person per generation, and we are no longer removing them.
Five generations have passed since the collapse of d began in earnest. Five clicks of a very slow ratchet. The Yamanashi mice made it to 58 clicks of a very fast ratchet before extinction. We’re not mice, and our ratchet isn’t turning as fast, because we still have recombination even if selection has been hobbled. But the direction of the ratchet is the same, and it doesn’t turn backwards on its own.
The cloning study’s authors noted, with considerable understatement, that their results “suggest that mammals rely on sexual rather than asexual reproduction to eliminate genetic anomalies caused by clonal reproduction.” What they should have said—what the math requires them to say, if they follow it to its conclusion—is that mammals rely on selection operating through sexual reproduction to maintain genomic integrity. Remove the sex, and you get Clone 58. Remove the selection, and you get a slower version of the same destination.
Sexual reproduction is necessary for maintaining the genome. But it is not sufficient. Selection is the engine. Sex is the transmission. You can have the finest gearbox ever engineered, but if the engine is dead, the car isn’t going anywhere.
The engine is dying. The Yamanashi experiment just showed us, in a twenty-year time-lapse, what happens at the end of that road, in approximately 1,150 years from 1900 Anno Domini.
The problem of the frozen gene is real. We had better pay attention.
DISCUSS ON SG