Pseudogenes in the Human Genome

What a pseudogene is, and why shared damage carries more weight than shared function, is set out in Pseudogenes, which covers the cases across birds, whales, bats, and penguins. This article applies the argument to the human genome.

Humans carry thousands of disabled genes. The evidential value is not that they are broken. It is that the same genes are broken in the same way, by the same mutations at the same positions, in exactly the species that common descent predicts should share them.

Shared function can be explained by shared design. Shared damage at identical locations is harder to account for that way.

Vitamin C#

Humans cannot synthesise vitamin C, which is why scurvy is a human disease and not a canine one. The gene for the final enzymatic step, GULO, is present in the human genome and does not work.

The guinea pig is the test case. Guinea pigs also require dietary vitamin C, so the functional consequence is identical to ours, and their gene is disabled differently — different deletions, different positions.

If the loss were a design decision for animals whose diet supplies the vitamin, there is no reason for primates to share one specific pattern of damage while guinea pigs carry another. If it is inherited, the primate lesions descend from a single ancestral event and the guinea pig lesion from a separate one. That is what is observed.

Egg yolk genes#

Egg-laying vertebrates produce vitellogenin, the protein that forms yolk. Placental mammals do not lay eggs and do not need it.

Humans carry recognisable remnants of three vitellogenin genes, flanked by the same neighbouring genes that surround them in the chicken genome — which is how the remnants were located.

The prediction ran in the useful direction here. The sequences were sought at a specific human chromosomal address because synteny with birds placed them there.

Human-specific losses#

Some disabled genes are broken in humans and functional in other apes, which dates the loss to the human lineage after its divergence.

The CMAH case is unusually well constrained. The inactivating deletion is present in all human populations and absent in all other apes, so it occurred after the human–chimpanzee split and before modern humans dispersed. The consequence is measurable in living people: human tissue lacks Neu5Gc, which is why immune reactions to it occur in recipients of some animal-derived medical products.

Olfactory receptor genes#

Humans have roughly 800 olfactory receptor genes, of which about 400 are pseudogenes — a higher proportion of broken to working than in mice or dogs, consistent with the reduced role of smell in primates that rely on vision.

Many of the disabling mutations are shared with other primates in the pattern the tree predicts, and the fraction broken increases as the lineage's reliance on vision increases.

Common objections#

"Pseudogenes have functions — 'junk DNA' was a failed prediction"

Some sequences classified as pseudogenes do have regulatory roles and several are transcribed. This is demonstrated and not disputed. Jonathan Wells's The Myth of Junk DNA (2011) and the Discovery Institute's use of ENCODE are addressed in Human Endogenous Retroviruses.

For the cases above the specific function is known and absent. Humans do not synthesise vitamin C, do not produce yolk protein, and do not make Neu5Gc. Whatever secondary role a sequence may play, the original function is gone and the deficit is measurable in the body.

The pattern would remain even if every sequence had some function. The evidence is that identical damage occurs at identical positions in exactly the species that would have inherited it.

"The designer reused a common blueprint, including inactive sections"

Possible, in the sense that any observation is compatible with unstated design intentions.

It predicts nothing. A blueprint reused with inactive sections could distribute those sections in any arrangement, so the observed one provides no support for the hypothesis.

Common descent predicts the arrangement in advance: damage nested according to the tree derived from other evidence, with species outside a group either retaining the working gene or having broken it differently. The guinea pig comparison is exactly that test, and it came out as predicted.

"Mutation hotspots could produce the same breaks independently"

Sequence context does affect mutation rates. CpG dinucleotides mutate more often and some regions are genuinely fragile.

Hotspots would produce recurrence at the same positions, which is a real prediction worth testing.

They do not produce the observed nesting. A hotspot mechanism should disable a gene repeatedly across many lineages independently, which is roughly what is seen in guinea pigs and in several bat lineages — different lineages, different lesions. What requires descent is the sharing: humans and chimpanzees carrying the same substitutions and the same single-base deletion, while a species with an identical dietary situation carries a different set.

"This assumes the gene was once functional rather than created broken"

The functional version exists in other animals and can be compared directly, so the comparison is not hypothetical.

The sequences retain the structure of working genes: recognisable exon boundaries, promoter remnants, and enough similarity for the disabling changes to be located precisely. A sequence created non-functional would not need those features.

Where an enzyme has been reconstructed by correcting the mutations, the repaired sequence produces a working protein — a direct test of whether the original was functional.

What the evidence shows#

Humans and other primates carry the same genes broken by the same mutations at the same positions, while animals with the same functional deficit outside the primate group carry different mutations.

Some losses are human-specific and can be dated to our own lineage — MYH16, CMAH, KRTHAP1 — with consequences still measurable in human physiology.

At least one pseudogene family, the vitellogenin remnants, was found by looking at the genomic address that synteny with birds predicted.

Common design accounts for shared working features. It does not predict shared damage distributed in a nested pattern matching the tree built from anatomy, retroviral insertions, and the fossil record.

The general treatment across animals is in Pseudogenes. Related human evidence is in Chromosome 2 and Human Vestigial Structures.