Human chromosomal fusion has often been cited as strong evidence supporting human-ape common ancestry. It’s not. As a recent paper in Cell Genomics reports, “Incomplete lineage sorting of segmental duplications defines the human chromosome 2 fusion site early during African great ape speciation.” It offers some new revelations about the supposed site of fusion in human chromosome 2.
We’ve written about this topic in the past (see here, here, and here for example). As I’ve said before, even if we fully grant that human chromosome 2 is indeed the result of the fusion of two chromosomes, all it shows is that somewhere within our human lineage a chromosomal fusion event occurred. Whether our lineage traces back to a common ancestor with apes is an entirely different question that isn’t settled by chromosomal fusion.
Paper Offers a Baroque Explanation
Now I’ve never felt strong motives to doubt human chromosomal fusion — if it turns out to be true, then I’m happy to accept it. Nonetheless there has been some curious evidence raised that calls it into question, at least to some degree. Fan et al. (2002) found that the supposed fusion site is “degenerated significantly” and “highly diverged from the prototypic telomeric repeats” compared to what you would expect from a pristine fusion site. This new paper in Cell Genomics tries to explain why the “fusion site” is so garbled, and tells a highly convoluted mutational tale to account for its current “degenerated” state. Here’s the short version of it:
[T]he fusion was associated with multiple pericentric inversions, segmental duplications (SDs), and the turnover of subterminal repetitive DNA.
This seems consistent with what we’ve heard before about the fusion site. I wrote about Fan et al. (2002) years ago here and noted that the paper even asked the question, “If the fusion occurred within the telomeric repeat arrays less than ~6 mya [million years ago], why are the arrays at the fusion site so degenerate?” It’s still a good question — does this new paper answer it?
Not in any elegant way. The story told in the new paper is awfully complicated and opaque. I’m not sure if it amounts to anything more than just a giant, baroque, after-the-fact 100 percent-evolution-assuming mutational just-so story. These kinds of stories are hard to refute but also make one wonder if they really are the correct answer, or if something else might be going on. All that aside, the paper does offer some fascinating new information about the supposed fusion site: it contains expressed genes that appear to be functional, and appears associated with important human neural functions.
Functional Genes at the Fusion Site
The paper reports that the fusion site may be functional and the authors support this claim with the finding that it contains four genes/pseudogenes that are expressed in various human tissues:
As the fusion site is nearly fixed in the human genome, we speculated that it may confer a functional advantage. Four putative noncoding genes/pseudogenes (PGM5P4, FAM138B, WASH2P, and DDX11L2) have been annotated at the site of the human chr2 fusion. According to GTEx short-read RNA sequencing (RNA-seq), these four genes/pseudogenes are expressed in testis, esophagus, fallopian tube, and cerebellum tissues. Further, both PGM5P4 and WASH2P are supported by long-read isoform sequencing (Iso-Seq) transcript data from CHM13hTERT and kidney tissue from ENCODE. In addition, methylation analysis demarcates a prominent CpG island showing the promoters/enhancers of PGM5P4, as identified using ONT reads from the T2T-CHM13 cell line.
Don’t miss the phrase that the fusion site “may confer a functional advantage” — the authors take its universal persistence in the human genome as evidence that it’s there because it’s doing something useful!
Potential Involvement with Neuronal Development
Even more curiously, the investigators then knocked out the fusion site in certain cells and found changes in “transcriptional programs involved in neural development”:
Differential gene expression analysis identified 547 upregulated and 869 downregulated genes using default settings in the differentially expressed gene (DEG) pipeline. … Functional enrichment analysis revealed that the upregulated genes were significantly associated with pattern specification processes (adjusted p = 8.61 × 10−5), whereas the downregulated genes were enriched for pathways related to neural development and organization, including forebrain development (adjusted p = 8.02 × 10−15) and axonogenesis (adjusted p = 4.92 × 10−14). These findings suggest that deletion of the fusion site may influence transcriptional programs involved in neural development, potentially contributing to phenotypic divergence between humans and NHPs.
If this fusion site has major effects upon transcription relating to neural development, that definitely does not sound like a random junk region resulting from the indiscriminate fusion of two chromosomes that then experienced all kinds of complex mutations over the last few million years. It seems like something far more important — something that may even show evidence of design.
Phylogenetic Conflicts
Another interesting finding from this paper is that when the supposed fusion site is compared to corresponding genomic sections in apes, there are patterns of similarity that contradict the standard phylogeny of great apes and humans. They found one region in the human fusion site that’s more similar to gorilla than it is to chimp or bonobo:
Of note, the SD_fusion_A (chr2) and other human copies share a monophyletic origin (∼3.3 mya), most closely related to gorilla copies ∼6 mya (95% confidence interval [CI]: 4.8–7.4 mya) instead of chimpanzee or bonobo — a pattern consistent with incomplete lineage sorting (ILS).”
Of course, the standard phylogeny claims that humans are more closely related to chimps than to gorillas, so this doesn’t fit the expected phylogenetic tree. They invoke incomplete lineage sorting (ILS) to explain this discrepancy, but that’s just an epicycle that we’ve frequently seen invoked by evolutionary biologists when some genetic data doesn’t fit the expected phylogenetic tree (see, for example, here). They attribute this quirky phylogenetic data to ILS, but ILS is at heart an explanation used to explain away phylogenetic incongruity.
They then invoke another instance of ILS (which really just means “phylogenetic incongruity”) to explain why another section of the genome is more similar in chimp/bonobo to gorilla than it is to human:
Similarly, SD_fusion_B is a 36 kbp segment corresponding to the partially truncated WASH245 (11 exons of an ancestral gene that potentially regulates actin cytoskeleton dynamics), FAM138B (3 exons of an ancestral gene of unknown function), and DDX11 (3 exons of an ancestral gene implicated in DNA metabolism). The segment exists as a single copy in both macaques and orangutans and originated from a locus mapping to ancestral African great ape chr12 (phylogenetic group XII). All human copies mapping to human chr2, chr20, chr12, and chr9 show a monophyletic origin (∼4.4 mya, 95% CI: 3.1–5.8 mya), suggesting human-specific duplications or interlocus gene conversion events. With respect to nonhuman African great apes, however, this clade diverges from a distinct monophyletic clade that includes chimpanzee, gorilla, and bonobo (∼7.4 mya, 95% CI: 5.4–9.3 mya). This topology is once again consistent with ILS.
Again, under the standard phylogeny, chimp and human are supposed to be closer to one another than they are to gorilla. So finding a section in the chimp/bonobo genome that is more similar to gorilla than to human is another example of phylogenetic incongruity. Not to worry — they just wave the magic wand of ILS to fix the problem.
All this supposed ILS makes them propose that the chromosomal fusion event occurred at a time when there was speciation going on way back in Australopithecus. That seems extremely speculative. What is ILS, after all? It’s a mechanism that is sometimes invoked to explain away gene sequences that don’t fit the standard phylogeny. Here the raw data shows evidence that is phylogenetically incongruent. They somehow turn that into evidence of a fusion event in Australopithecus.
What Has This Paper Demonstrated?
In the end, what this paper actually seems to demonstrate is
- A convoluted tale involving a series of complicated mutations is necessary to answer Fan et al. (2002)’s question about why the “fusion site” looks so “degenerate” compared to what you would expect a “fusion site” to look like. It’s not exactly an elegant or compelling explanation and it leaves one wondering if something else is going on here.
- The “fusion site” seems to contain multiple expressed functional genes, including sections that have major effects upon transcription related to neural development. This is not exactly what you would expect from a site that is the result of the random fusion event that then experienced all kinds of complex mutations!
- The fusion site has various subsections that are distributed among humans and apes in patterns that contradict the standard phylogeny.
The paper predictably invokes ILS to explain away these phylogenetic conflicts — even claiming this is evidence the fusion event occurred during speciation in the genus Australopithecus. Perhaps, but overall there’s a lot of data here that is certainly not what you’d expect from an evolutionary perspective.