Rapidly jumping genes from mitochondria to nuclear chromosomes show how dynamic our cell’s genomes are, which might have implications for how long we live.
For decades we treated the nuclear genome as the “master script” and mitochondrial DNA (mtDNA) as a small, mostly static side note. But accumulating data suggest mtDNA variation and signaling aren’t passive; they can influence nuclear gene expression, stress responses, metabolism, and even inflammatory tone through retrograde signaling.
A few implications that make this exciting from a systems biology perspective:
1. Genome is not just sequence, but it’s cross-talk. Nuclear and mitochondrial genomes are in constant bidirectional communication. mtDNA variants can shift metabolic flux, redox balance, and epigenetic regulation in ways that alter nuclear transcriptional programs.
2. Energy state as information. Mitochondria aren’t just ATP factories; they’re sensors. Changes in mitochondrial function can reshape chromatin states and stress signaling pathways, effectively modulating how the nuclear genome is “read.”
3. Heteroplasmy matters. The proportion of different mtDNA variants within cells can influence phenotype, and that distribution can drift with age, stress, and tissue context, introducing another dynamic layer to genetic expression over the lifespan.
4. Disease and aging connections. If mtDNA can reprogram nuclear responses, it may help explain variability in aging trajectories, stress resilience, and susceptibility to metabolic or neurodegenerative disease.
What I appreciate most is that this shifts the narrative from “genes are destiny” to “genes are part of a responsive bioenergetic network.” It opens up a more dynamic view of biology, where environment, stress, and metabolism interface directly with genetic expression.
Wow. Fascinating. Thanks for sharing, and also for the work you do. Given the “Maternal” nature of Mitochondria, I can’t help but also wonder if, instead of this process causing accelerated aging in response to stress, it might be an attempt to help prevent accelerated aging in response to stress? In other words, does the mere presence of more Numts show that they seem to be somehow correlated to aging, or does it seem more certain that they seem negatively correlated to health and aging? (Is a correlation to increased aging considered positive or negative lol 😆?). 🤔
Good question. The correlation we found shows that people with more Numts in their brain are those who died earlier. The older the people when they dies, the less Numts they had.
So they are not a marker of age, but may be a marker that the biology needs to "evolve" or transform faster (to adapt to some stressor, perhaps?).
But there is a cost and price to pay to adapt--it costs energy, and therefore the more you have to do, the more you take away from other resources that keeps you healthy and alive. This is speculative but aligns with the energetic costs of allostasis/adaptation:
This piece reminds me of the words of Marcus Aurelius: “The mind is that which is roused and directed by itself. It makes of itself what it chooses. It makes what it chooses of its own experience.”
I know this is going to sound a little crazy, but these findings on numtogenesis map strikingly onto the transformer architecture that dominates machine learning.
Mitochondria broadcast mtDNA into the nucleus the way attention heads broadcast queries into a shared representational space. Numts that rewrite the nuclear genome parallel how each attention layer rewrites the residual stream that all subsequent layers inherit.
The prefrontal cortex accumulates more Numts than the cerebellum, just as deeper transformer layers accumulate more modifications to the residual. Most provocatively, your finding that stress accelerates numtogenesis mirrors how noisy inputs force more aggressive coarse-graining in neural networks, sometimes destructively. The candle flame is a forward pass.
The engine analogy doesn't work because the information encoded in our cells is subject to multiplicative, not additive noise.
I'm feeling my way through this, but I'm beginning to develop the impression that the transformer is a discrete transistor-driven implementation of a free energy minimizing architecture that works across many scales. It feels like a microcosm for evolution itself.
Joseph McCard distils Martin’s laboratory findings down to what are, for him, cohesive findings. Layman that I am, I must distil further. I gather from the findings of both Martin and Joseph the following, and beg for corrections: Action (including stress) affects identity. Identity (which is fluid) determines longevity. Our energy flow is our self, which in turn drives our apparatus onward...or doesn’t. I don’t mean to sound poetic, on the contrary, but I need to translate the stuff into usable English.
Great question! The new Insertions we observed, were enriched in areas of open chromatin.
Large scale data in tens of thousands of people show that there are some hotspots for insertions at evolutionary scale, with some rare examples of insertions in regulatory or coding elements driving disease.
This is a fascinating and important reframing!
For decades we treated the nuclear genome as the “master script” and mitochondrial DNA (mtDNA) as a small, mostly static side note. But accumulating data suggest mtDNA variation and signaling aren’t passive; they can influence nuclear gene expression, stress responses, metabolism, and even inflammatory tone through retrograde signaling.
A few implications that make this exciting from a systems biology perspective:
1. Genome is not just sequence, but it’s cross-talk. Nuclear and mitochondrial genomes are in constant bidirectional communication. mtDNA variants can shift metabolic flux, redox balance, and epigenetic regulation in ways that alter nuclear transcriptional programs.
2. Energy state as information. Mitochondria aren’t just ATP factories; they’re sensors. Changes in mitochondrial function can reshape chromatin states and stress signaling pathways, effectively modulating how the nuclear genome is “read.”
3. Heteroplasmy matters. The proportion of different mtDNA variants within cells can influence phenotype, and that distribution can drift with age, stress, and tissue context, introducing another dynamic layer to genetic expression over the lifespan.
4. Disease and aging connections. If mtDNA can reprogram nuclear responses, it may help explain variability in aging trajectories, stress resilience, and susceptibility to metabolic or neurodegenerative disease.
What I appreciate most is that this shifts the narrative from “genes are destiny” to “genes are part of a responsive bioenergetic network.” It opens up a more dynamic view of biology, where environment, stress, and metabolism interface directly with genetic expression.
Really compelling synthesis!
Great summary!
It is wild to think of our brains as a living diary where stress leaves actual genetic footprints. Thanks for the great article.
Agreed. So many different kinds of energetic footprints. Energy flow through a living system always leaves traces.
Wow. Fascinating. Thanks for sharing, and also for the work you do. Given the “Maternal” nature of Mitochondria, I can’t help but also wonder if, instead of this process causing accelerated aging in response to stress, it might be an attempt to help prevent accelerated aging in response to stress? In other words, does the mere presence of more Numts show that they seem to be somehow correlated to aging, or does it seem more certain that they seem negatively correlated to health and aging? (Is a correlation to increased aging considered positive or negative lol 😆?). 🤔
Good question. The correlation we found shows that people with more Numts in their brain are those who died earlier. The older the people when they dies, the less Numts they had.
So they are not a marker of age, but may be a marker that the biology needs to "evolve" or transform faster (to adapt to some stressor, perhaps?).
But there is a cost and price to pay to adapt--it costs energy, and therefore the more you have to do, the more you take away from other resources that keeps you healthy and alive. This is speculative but aligns with the energetic costs of allostasis/adaptation:
https://www.sciencedirect.com/science/article/pii/S030645302200292X?via%3Dihub
Interesting questions Scott.
If you were to ask me, I would say that a positive correlation would exist between ⬆️ Numts and an increase in what we call aging.
Is there a chicken and egg situation here between aging and stress or is this a false dichotomy?
Does aging cause the stress response leading to more Numts-o-genesis or is it vice versa?
I actually just wrote an article about a related phenomenon!
https://drewdespereaux.substack.com/p/how-cancer-hijacks-your-immune-system?utm_source=share&utm_medium=android&r=63rclf
This piece reminds me of the words of Marcus Aurelius: “The mind is that which is roused and directed by itself. It makes of itself what it chooses. It makes what it chooses of its own experience.”
— Meditations 6.8.
I know this is going to sound a little crazy, but these findings on numtogenesis map strikingly onto the transformer architecture that dominates machine learning.
Mitochondria broadcast mtDNA into the nucleus the way attention heads broadcast queries into a shared representational space. Numts that rewrite the nuclear genome parallel how each attention layer rewrites the residual stream that all subsequent layers inherit.
The prefrontal cortex accumulates more Numts than the cerebellum, just as deeper transformer layers accumulate more modifications to the residual. Most provocatively, your finding that stress accelerates numtogenesis mirrors how noisy inputs force more aggressive coarse-graining in neural networks, sometimes destructively. The candle flame is a forward pass.
The engine analogy doesn't work because the information encoded in our cells is subject to multiplicative, not additive noise.
Very nice analogy, Michael.
Here's a more complete mapping.
https://substack.com/@michaelfrankmartin280311/note/c-217737100?r=4ylj5&utm_source=notes-share-action&utm_medium=web
I'm feeling my way through this, but I'm beginning to develop the impression that the transformer is a discrete transistor-driven implementation of a free energy minimizing architecture that works across many scales. It feels like a microcosm for evolution itself.
Joseph McCard distils Martin’s laboratory findings down to what are, for him, cohesive findings. Layman that I am, I must distil further. I gather from the findings of both Martin and Joseph the following, and beg for corrections: Action (including stress) affects identity. Identity (which is fluid) determines longevity. Our energy flow is our self, which in turn drives our apparatus onward...or doesn’t. I don’t mean to sound poetic, on the contrary, but I need to translate the stuff into usable English.
Accurate - thank you for the distillation.
This is the main thesis of the ENERGY book.
OK, so just chop off their frontal lobes so they live longer.
Psychiatry tried a version of this a few decades ago.
Great question! The new Insertions we observed, were enriched in areas of open chromatin.
Large scale data in tens of thousands of people show that there are some hotspots for insertions at evolutionary scale, with some rare examples of insertions in regulatory or coding elements driving disease.