Hunting ancient mitochondrial scraps
that evolution may have recycled
Jayden · Frith Lab · University of Tokyo · UTSIP 2026
If two species share a feature, they inherited it from a shared ancestor.
So the feature is at least as old as that ancestor.
Tree and dates from TimeTree, across 16 mammal genomes
A piece of the mitochondria's own DNA is copied into the nucleus.
That copy is a NUMT, and most slowly fall apart as molecular fossils.
NUMT means nuclear mitochondrial DNA segment
We compare the human NUMT to many other animals.
The more distant the animal that still shares it, the older it must be.
We found 189 ancient NUMTs, and 90 are older than 100 million years.
LAST, comparing DNA to DNA and DNA to protein
For each NUMT I check which functional features it overlaps in the genome.
Each one is a clue that the old scrap might now be doing a job.
It sits in an exon, so it ends up in a sentence the cell actually reads out.
It overlaps a regulatory element, like a switch that turns a gene on or marks where reading begins.
The DNA here is unpacked and proteins land on it, like a well thumbed page.
We find RNA copies of this stretch, so it really is being read, not just sitting there.
Repeats, duplicated regions, and hard to map DNA can fake a match, so these lose points.
Conservation is left out on purpose. For a NUMT it is misleading, because the nuclear copy can look conserved just by resembling the still working mitochondrial gene.
Add up the features, subtract the doubts, and sort the 189.
Watch which one keeps rising to the top.
An exploratory ranking to prioritise, not a measure of significance
Out of 189 NUMTs, only one sits inside protein coding DNA.
That is the one worth opening up.
chr19:49,854,479 to 49,854,646 · 167 bp · a brand new find that sits inside a gene
The red bar is our NUMT, sitting inside a PTOV1 gene exon
The red part is NUMT derived, sitting at the edge of a working protein region
In an ancestor of all placental mammals, a scrap of mitochondrial DNA slips into the nucleus.
Instead of decaying, it is kept and built into a gene we now call PTOV1, becoming part of one of its exons. Evolution recycled it.
PTOV1 is active in our cells and is linked to cancer, where it is over expressed in prostate and other tumours.
So a 100 million year old scrap of mitochondria, that looks irrelevant to modern humans,
might be woven into a gene that matters to our health. That is the idea we now want to test.
Built a pipeline and classified the function potential of all 189 ancient NUMTs, drawing gene, regulatory, and conservation evidence into one exploratory ranking.
Review the literature on the top candidates, and study the most interesting region in depth, the PTOV1 NUMT, to ask whether it is really a working part of the protein.
The PTOV1 NUMT at a glance
Jayden · Frith Lab · University of Tokyo · UTSIP 2026