Macrauchenia: the good and bad of genomic studies

From the Wesbury et al. 2021 abstract
“The unusual mix of morphological traits displayed by extinct South American native ungulates (SANUs) confounded both Charles Darwin, who first discovered them, and Richard Owen, who tried to resolve their relationships. Here we report an almost complete mitochondrial genome for the litoptern Macrauchenia (Fig. 1). Our dated phylogenetic tree (Fig. 2) places Macrauchenia as sister to Perissodactyla, but close to the radiation of major lineages within Laurasiatheria. This position is consistent with a divergence estimate of B66Ma.”

Note they don’t ask us to pay as much attention to the proximal outgroup for Macrauchenia: the clade Carnivora (Fig. 2).

Figure 1. Macrauchenia museum mount.

Figure 1. Macrauchenia museum mount.

According to Wikipedia
Laurasiatheria is a gene-based clade “that includes that includes hedgehogs, even-toed ungulates, whales, bats, odd-toed ungulates, pangolins, and carnivorans, among others.”

Isn’t that an odd assemblage? 
Think about it. According to Wesley et al. (Fig. 2), sabertooth cats are closer to horses, rhinos and Macrauchenia than other long-legged, placental herbivores. By the way, in gene studies elephants appear in an unrelated major clade, Afrotheria.

Figure 1. Gene-based cladogram from Westbury et al. 2021 (slightly compressed to fit). Note the close relationship between Carnivora and Macrauchenia here. That is not replicated in a trait-based study (Fig. 2).

Figure 2. Gene-based cladogram from Westbury et al. 2021 (slightly compressed to fit). Note the close relationship between Carnivora and Macrauchenia here. That is not replicated in a trait-based study (Fig. 2).

A more reasonable, trait-based, phylogenetic analysis
(the large reptile tree, LRT, 1794+ taxa, subset Fig. 3) also nests the Macrauchenia clade basal to tapirs, rhinos and horses. The outgroup is the hyrax + elephant + manatee clade, then the artiodactyls, then the mesonychids + hippos + desmostylians + mysticetes. Off this chart (Fig. 3), the clade Carnivora is the basalmost placental clade, not the proximal outgroup to Macrauchenia.

Figure 2. Subset of the LRT focusing on derived placentals. Yellow highlights the Macrauchenia clade.

Figure 3. Subset of the LRT focusing on derived placentals. Yellow highlights the Macrauchenia clade.

Perhaps taxon exclusion is at fault here.
On the other hand, gene studies too often produce such odd interrelationships (Carnivora nesting closer to Macrauchenia than other herbivore clades). Gene studies too often deliver false positives in deep time studies. That’s a fact, not a hypothesis.

If your professor is asking you to help out on a deep time genomic study,
run.


References
Westbury M et al. (21 co-authors) 2021. A mitogenomic timetree for Darwin’s enigmatic South American mammal Macrauchenia patachonica. Nature Communications | 8:15951 | DOI: 10.1038/ncomms15951 |www.nature.com/naturecommunications

https://en.wikipedia.org/wiki/Laurasiatheria
reptileevolution.com/macrauchenia.htm

Woltering et al. 2020 study genes to elucidate finger origins

Woltering et al. 2020
attempted to elucidate the transition from fins to fingers by studying the genes of extant lungfish, which don’t have fingers and their ancestors never had fingers.

From the abstract
“How the hand and digits originated from fish fins during the Devonian fin-to-limb transition remains unsolved.

No. The large reptile tree (LRT; subset Fig. 1) solved that problem in 2019 following the work of Boisvert, Mark-Kurik and Ahlberg 2008. These authors found four finger buds on Panderichthys. Thereafter four fingers appear on all basalmost tetrapods in the LRT, like Trypanognathus (Fig. 2), a taxon found in Carboniferous strata with Middle Devonian origins. Taxon exclusion is once again the problem here.

Late Devonian taxa with supernumerary digits, like Acanthostega and Ichthyostega, are the traditional ‘go-to’ taxa for the fin-to-finger transition. That was supplanted in 2019 by phylogenetic analysis in the LRT (subset Fig. 1). Simply adding taxa recovers Acanthostega and Ichythyostega as terminal taxa. They have more derived skulls and bodies sporting larger limbs and more digits.

Figure 4. Subset of the LRT focusing on basal tetrapods. Colors indicate number of fingers known. Many taxa do not preserve manual digits.

Figure 4. Subset of the LRT focusing on basal tetrapods. Colors indicate number of fingers known. Many taxa do not preserve manual digits.

Woltering et al. 2020 report,
“Controversy in this conundrum stems from the scarcity of ontogenetic data from extant lobe-finned fishes. We report the patterning of an autopod-like domain by hoxa13 during fin development of the Australian lungfish, the most closely related extant fish relative of tetrapods.”

In other words, Woltering et al. looked at genes in lungfish that never had digits.

Figure 6. Dorsal and ventral views of Panderichthys and several basal tetrapods demonstrating the low, flat skulls and bodies with small limbs and relatively straight ribs.

Figure 2. Dorsal and ventral views of Panderichthys and several basal tetrapods demonstrating the low, flat skulls and bodies with small limbs and relatively straight ribs at the fin-to-finger transition. Acanthostega and Ichythyostega have more derived bodies with larger limbs and more digits.

Why study lungfish
when we have fossil taxa (Fig. 3) in the lineage of tetrapods? Why study genes when genomic studies produce false positives in deep time? Taken together the Woltering et al. study seems like a waste of effort on both fronts, but they didn’t realize this at the time. Paleontologists love genomics like Isaac Newton loved alchemy.

Figure 3. Forelimb of several basal tetrapods rearranged to more closely fit the LRT. Four fingers turns out to be the primitive number. Five is a recent mutation. Six was a short-lived experiment in Tulerpeton.

Figure 3. Forelimb of several basal tetrapods rearranged to more closely fit the LRT. Four fingers turns out to be the primitive number. Five is a recent mutation. Six was a short-lived experiment in Tulerpeton.

Woltering et al. 2020 report,”
“Differences from tetrapod limbs include the absence of digit-specific expansion of hoxd13 and hand2 and distal limitation of alx4 and pax9, which potentially evolved through an enhanced response to shh signaling in limbs. These developmental patterns indicate that the digit program originated in postaxial fin radials and later expanded anteriorly inside of a preexisting autopod-like domain during the evolution of limbs. Our findings provide a genetic framework for the transition of fins into limbs that supports the significance of classical models proposing a bending of the tetrapod metapterygial axis.”

Be wary of genetic studies over deep time. They have been shown to deliver false positives way too often to be trusted, or even attempted. Fossils and phenomic studies are better in all respects because they recover cladograms in which all taxa demonstrate a gradual accumulation of derived traits.


References
Boisvert CA, Mark-Kurik E and Ahlberg PE 2008.
 The pectoral fin of Panderichthys and the origin of digits. Nature 456:636–638.
Woltering JM et al. (5-co-authors) 2020. Sarcopterygian fin ontogeny elucidates the origin of hands with digits. Science Advances 6(34): eabc3510 DOI: 10.1126/sciadv.abc3510
https://advances.sciencemag.org/content/6/34/eabc3510

Kuhl et al. 2020: an ‘unbiased, resolved’ avian tree of life?

Similar to other traditional myths,
Kuhl et al. 2020 repeat a traditional myth that will never be validated.

From the abstract
“In this tree, grebes and flamingos are the sister clade of all other Neoaves,”

Kuhl et al. is a genomic study. These two bird types do not resemble each other (Fig. 1) and do not nest together in phenomic studies, like the large reptile tree (LRT, Subset Fig. 2). Prum et al. 2015 recovered the same false positive relationship. If a study cannot be validated by traits, it is invalid and worthless.

The unanswered question here is, “Why do certain genes resemble one another in these two unrelated taxa?” Now there’s another subject for a PhD thesis.

Figure 2. Flamingo and grebe illustration from Nat Geo article on birds.

Figure 1. Flamingo and grebe illustration from Nat Geo article on birds.

Earlier we looked at grebes and flamingoes
here in 2017. Apparently there has been no enlightenment among paleontologists since then. Where is the critical thinking?

Continuing from the Kuhl et al. abstract:
“All non-passerine taxa were placed with robust statistical support including the long-time enigmatic hoatzin (Opisthocomiformes), which was found being the sister taxon of the Caprimulgiformes.” [= nightjars, sailors [= barn swallows?], hummingbirds, unrelated taxa in the LRT]

In the LRT the hoatzin nests between sparrows and parrots because they look alike, both overall, and in detail, location and diet. Relatives

If you have not learned this already,
genomic studies over deep time lead one to madness.

From the press release:
“For the first time, they have been able to clarify the relationship of all families of non-passerine birds and almost all families of passerine birds.  The new family tree is based on gene sections that do not code for proteins, but contain sequences that are specific to the families and their genera.”

Better to stick with a phenomic study that makes sense, like the LRT. Chickens and ducks are not related in the LRT.  That makes sense because they don’t resemble each other. Parrots and hawks are not related in the LRT. That also makes sense, but genomic studies put each of these two pairs of taxa together.

As mentioned above, we looked at Prum et al 2015
earlier in a three-part series ending here demonstrating several times that genomic studies do not replicate phenomic studies in the sort of deep time that brought us birds (Late Jurassic to the Present).

Genomic studies are EXCELLENT
over short phylogenetic time. For this reason genomic studies are used by the judicial system. For this reason paleontologists think genomic studies should supersede phenomic  studies. However, tests document the fact that genomic studies produce mismatches (taxa that do not look alike, Fig. 1) over deep time. Paleontologists realize this, but continue to put their faith in genomic studies, hoping that someday the last common ancestor of flamingoes (genus: Phoenicopterus) and grebes (genus: Aechmophorus) will someday be discovered in the fossil record.

The LRT has already recovered that last common ancestor,
a taxon close to the late-surviving, but extant megapodes, Megapodius and Early Cretaceous Juehuaornis. You’ll notice (Fig. 1) that clade includes nearly all crown birds.

Figure 3. Subset of the LRT focusing on birds. Note the separation of the duck clade from the chicken clade.

Figure 1. Subset of the LRT focusing on birds. Note the separation of the duck clade from the chicken clade.

The hallmarks of the Scientific Method include:

  1. asking a specific question
  2. devising a hypothesis
  3. experimenting to gather data
  4. analyzing the data, and then
  5. evaluating whether the hypothesis is correct based on the experimental data.

When the data support the hypothesis, the findings can be published or shared.
What happens if the findings do not confirm the hypothesis? Answers here.

Sadly, 
Kuhl et al. 2020 was a complete waste of time for the nine co-authors, the editors, referees and innumerable readers who all believed the flawed premise of this genomic hypothesis of interrelations. Where is the genomic study that replicates phenomic studies like the LRT? When that happens, we’ll have that long sought validation.


Pertinent Carl Sagan – ‘A way of thinking’ YouTube video.

References
Kuhl H, et al. (8 co-authors) 2020. An unbiased molecular approach using 3″UTRs resolves the avian family-level tree of life. Molecular Biology and Evolution, msaa191 (advance online publication) doi: https://doi.org/10.1093/molbev/msaa191
https://academic.oup.com/mbe/advance-article/doi/10.1093/molbev/msaa191/5891114

Snake origin paper fails due to genomics and fakes an ancestral snake skull

Watanabe et al. 2019
“demonstrate that highly diverse phenotypes, exemplified by lizards and snakes, can and do arise from differential selection acting on conserved patterns of phenotypic integration.”

To build their phylogenetic tree, Watanabe 2019 report,
“To conduct comparative phylogenetic methods, we constructed a time-calibrated phylogenetic tree by using a published time calibrated molecular phylogeny of extant squamates (Zheng and Weins 2016) and incorporating extinct taxa based on previous systematic work and fossil occurrence data from the Paleobiology Database (paleobiodb.org). We grafted extinct taxa onto the extant tree by applying the equal-branching method based on the mean of first occurrence age range. Although the phylogenetic placement of Mosasauria within squamates remains ambiguous, Plotosaurus was placed within the molecular phylogeny as a sister taxon to Serpentes (“Pythonomorpha hypothesis”) and Polyglyphanodon as sister to iguanians in accordance with the phylogeny based on combined molecular and morphological data.”

Taxon exclusion is once again the problem here.
Too few fossil taxa appea in the Watanabe et al. cladogram. The keyword ‘outgroup’ is not found in the text or SuppData. Watanabe et al. report, Sphenodon was not included in analyses with the exception of morphospaces.” As a result the highly derived legless amphisbaenid, Dibamus, was chosen as the outgroup. That is wrong, according to the large reptile tree (LRT, 1524 taxa) in which iguanids are plesiomorphic basalmost squamates. Sphenodontids and tritosaurs are outgroups in the LRT, are not represented.

Mosasaurs, like Plotosaurus, arise from varanids in the LRT. The clade Pythonomorpha is not recovered by the LRT.

Snake ancestors (none of which appear in Watanabe et al.) arise from Jurassic gekko-like ancestors, like Tchingisaurus and ending with tiny Tetraphodophis (Fig. 1). The list of snake ancestors includes terrestrial taxa, like Ardeosaurus, and aquatic taxa, like Pontosaurus. Fossorial (burrowing) snakes arise as derived taxa, not primitive forms. 

Figure 1. Hypothetical ancestral snake skull compared to real ancestral snake skull.

Figure 1. Hypothetical ancestral snake skull compared to real ancestral snake skull. Watanabe et al. 2019 should have stayed away from fiction.

Quoting the NHM.AC.UK press release,
“Prof Goswami says, ‘There is a lot of debate about how snakes evolved, but we think we have traced the ancestral skull shape. Lots of scientists have speculated that maybe snake ancestors lived in water, which made them lose their legs. So it’s surprising that the patterns we saw led us to a semi-fossorial animal.’” 

There is no debate
when you use the last common ancestor approach based on phenomic (trait-based) phylogenetic analyses. Following in the footsteps of those who imagined pterosaur ancestors, the Watanabe team imagined a snake ancestor for no reason, because the LRT provides a long list of snake ancestors going back to Silurian jawless fish.

Gene studies
work well in criminal identification closely related taxa within a genus. Genomics fail for various reasons in deep time studies. This is something paleontologists and biologists need to realize. If you want results in which all derived taxa demonstrate gradually accumulating traits, you have to use trait-based studies and fossil taxa. Why turn your back on proven results in favor of a method you hope works, but never does?


References
Watanabe A, Fabre A-C, Felice RN, Maisano JA, Müller J, Herrel A and Goswami A 2019. Ecomorphological diversification in squamates from conserved pattern of cranial integration. www.pnas.org/cgi/doi/10.1073/pnas.1820967116
Zheng Y and Wiens JJ and 2016. Combining phylogenomic and supermatrix approaches, and a time-calibrated phylogeny for squamate reptiles (lizards and snakes) based on 52 genes and 4162 species. Mol. Phylogenet. Evol. 94, 537–547.

National History Museum News

 

 

New passerine genomic study not confirmed by phenomic study

Oliveros et al. 2019
produced an exhaustive DNA study from 137 passerine families, then calibrated their phylogeny using 13 fossils to examine the effects of different events in Earth history on the timing and rate of passerine diversification.

Unfortunately
the large reptile tree (LRT, 1434 taxa) produced a different tree because it uses phenomic traits, not genes.

The two trees both started with birds of prey, including owls.
Then they diverged. The Oliveros team recovered 137 families of passerines arising from highly derived parrots, arising from highly derived owls.

The LRT recovered highly derived parrots arising from the more primitive hoatzin Opisthocomus, arising from the more primitive sparrow, Passer, arising from the more primitive grouse + chickens + peafowl and kin going back to Early Cretaceous Eogranivora. In the LRT owls give rise to birds of smaller prey: owlets, like Aegotheles, and swifts, like Apus, not herbivorous parrots.

Figure 1. Skeleton of the common house sparrow, Passer domestics.

Figure 1. Skeleton of the common house sparrow, Passer domestics. Note the heavy, seed-crunching beak, a precursor for the heavier see-crunching beak in parrots, not the other way around.

Among the traditional ‘passerines’ tested by the Oliveros team
are the distinctively different crows (genus Corvus) and nuthatches (genus Sitta). These clades nest apart from each other in the LRT and apart from Passer, the sparrow. In the LRT, crows and nuthatches are not Passerines, but parrots and hoatzins are passerines. Sometimes competing cladograms can be topsy-turvy like that, with similar sister taxa flipped with regard to primitive and derived. Earlier I mentioned ‘woodpeckers’, which have never been considered passerines, because woodpeckers and nuthatches are sisters in the LRT.

Robins (genus: Turdus) are considered passerines in the DNA study. They are crow relatives in the LRT. Jays (genus: Cyanocitta) and grackles (genus: Quiscalus) are crow relatives in the LRT. Neither are included in the DNA study that includes crows (genus: Corvus).

Figure 1. Several birds with zygodactyl feet (light red) and one member of the clade Zygodactylidae (red).

Figure 2. Subset of the LRT focusing on birds. This is how they are related to one another based on phenomic traits. Note the presence of Passer nesting between the chicken, Gallus and the parrot, Ara. Other purported passerines, like Turdus, Corvus and Sitta,  nest in other clades here.

So, once again,
when taxonomists use genomic (DNA) tests they run the risk of wasting their time when dealing with deep time taxa. Some paleo and bird workers put their faith in DNA, hoping it will recover relationships because it works well in humans. Unfortunately, too often phenomic tests are at odds with genomic tests to put  faith in genomic tests. Only phenomic (trait) tests recover cladograms that produce a gradual accumulation of traits among sister taxa, echoing deep time events. Only phenomic tests can employ fossils. Let’s not forget our fossils.

A suggestion for Oliveros et al. 2019:
test your results against your own phenomic study. If valid, both of your results will be the same. If not, one of your tests needs to be trashed.


References
Oliveros CH and 31 co-authors 2019. Earth history and the passerine superradiation.

www.pnas.org/cgi/doi/10.1073/pnas.1813206116