Recalibrating clade origins, part 4

Earlier
we looked at the first part and second part and third part of Marjanovic’s 2019 chronological recalibration of vertebrate nodes. Today we continue in part 4 of 5.

Mammalia (Prototheria + Theria)
Based on the literature, Marjanovic 2019 considers morganucodontans and tiny Hadrocodium to be mammalomorphs, not mammals. He is unsure about haramiyidans. With regard to the first dichotomy of mammals, he reported, “I recommend a hard minimum age of 179 Ma for this calibration.”

By contrast the large reptile tree (LRT, 1630+ taxa) nests Morganucodon (Late Triassic, 205 mya), Hadrocodium (Early Jurassic) and Henosferus (Middle Jurassic) together in the most basal subclass within Theria, part of the first dichotomy within Mammalia. Marjanovic considered Henosferus one of the oldest uncontroversial mammals at 179 mya. Megazostrodon is a late surviving (early Jurassic) last common ancestor taxon of all mammals in the LRT. It must have appeared prior to Morganucodon in the Late Triassic.

Theria (Metatheria + Eutheria)
Marjanovic reports, “The oldest securely dated eutherian is Ambolestes at 126 Ma.” Then he reports, “Accepting that Juramaia is not from the Lanqi Fm, I propose 160 Ma as the soft maximum age of this calibration.”

By contrast, Morganucodon (Late Triassic, 210mya) nests as the oldest therian. Ambolestes nests with Didelphis, the opossum, within the Theria, not Eutheria.  Thereafter the traditional Metatheria splits in three clades in the LRT, a largely herbivorous branch with Glironia and Marmosops at its base, and a largely carnivorous branch with Monodelphis and Chironectes at the base of one branch and Caluromys + Placentalia at yet another. So, while Caluromys (Fig. 3) retains a pouch, it is also the last common ancestor of all placentals.

Figure 1. Pteropus and Caluromys compared in vivo and three views of their skulls. Caluromys is in the ancestry of bats and shows where they inherited their inverted posture.

Figure 1. Pteropus and Caluromys compared in vivo and three views of their skulls. Caluromys is in the ancestry of bats and shows where they inherited their inverted posture.

Marjanovic often errors by not including extant taxa that are more primitive than extinct taxa that are older. This comes back to bite him several times, especially so when he relies on a single fossil tooth rather than a living animal he can hold. The LRT tests both living and extinct taxa to minimize taxon exclusion.

Marjanovic discusses the possibility that Sinodelphys is the oldest known metatherian, but Sinodelphys nests as one of the most primitive prototherians in the LRT, as we learned earlier here.

Placentalia (Atlantogenata + Boreo(eu)theria)
In the world of gene studies, Atlantogenata include the highly derived elephants and anteaters. The Boreoeutheria include the highly derived whales, humans and hooved mammals. Genomic studies deliver false positives, and these are among the most blatant, so ignore these. They don’t deliver a gradual accumulation of derived traits.

By contrast, in the LRT the first dichotomy in the placentalia splits arboreal Vulpavus from arboreal Nandinia and thereafter arboreal Carnivora (mongooses and raccoons) from arboreal Volantia (bats and colugos) + arboreal Primates and the rest of the Placentalia. All of these civet-like and tree opossum-like taxa look like Caluromys (Fig. 1), as you can see. Elephants and anteaters come later. Adding living taxa to Marjanovic’s search for primitive placentals would have helped clarify his research and conclusions, preventing him form perpetuating old myths.

Carnivora (Feliformia + Caniformia)
Marjanovic errs by reporting the basal dichotomy within Carnivora splits cats from dogs.

By contrast in the LRT cats and dogs are closely related and derived taxa, not basal. As mentioned above, civets, mongooses and raccoons are basal Carnivora.

Euarchontoglires/Supraprimates (Gliriformes + Primatomorpha)
Marjanovic discusses several poorly preserved, sometimes one tooth only, fossil taxa from the early Paleocene (65mya). Some of these are anagalids, which nest at the base of yet another clade in the LRT, the one with tenrecs and odontocetes (toothed whales).

By contrast in the LRT lemur-like adapids appear at the base of the Primates. Tree-shrews appear at the base of the Glires.

Marsupialia (Didelphimorphia – Paucituberculata + Australidelphia)

  • Didelphimorphia = opossums from North America
  • Paucituberculata = South American marsupials, sans Dromiciops
  • Australidelphia = Australian and Asian marsupials, plus Dromiciops

Marjanovic reports, “I therefore propose 55 Ma as a probably overly strict hard minimum age for this calibration.” He later reports, “Rather than the beginning of the Maastrichtian, I propose the beginning of deposition of the Lance and Hell Creek formations, where Glasbius has been found, as the hard maximum age for this calibration, which I estimate as 68 Ma.”

See Figure 1 for a different three-part marsupial split from the LRT. Dromiciops is only one of many similar herbivorous marsupials. Middle Late Cretaceous Asioryctes is a basal member of the largely herbivorous clade. Early Cretaceous Vincelestes is a basal member of the largely carnivorous clade. So Middle to Late Jurassic (175mya) is a better estimate for the genesis of marsupial diversity. That means marsupials dispersed during the Pangean era without the need of an oceanic dispersal.

Marjanovic mistakenly reports, “Marsupials, other metatherians and indeed other therians are wholly absent from the Late Cretaceous mammaliform record of South America, which consists instead of gondwanatherian haramiyidans and a very wide variety of meridiolestidan stem-theriiforms.”

  • Meridiolestida = non-therian mammals (= Prototheria, Montremata) seems to be based on tooth traits. Cronopio and Necrolestes are among the only tested taxa also  found in the LRT. Cronopio is an omnivorous member of the pre-metatherian Theria in the LRT. Necrolestes is a basal member of the placental clade, Glires, derived from the treeshrew Tupaia in the LRT. So, again, we have a mismatch due to not testing all the mammals against all the mammals. That is what makes the LRT such a powerful tool that should be more widely used to avoid such old school mythology.

More tomorrow as we conclude part 5 of 5.


References
Marjanovic D 2019. Recalibrating the transcriptomic timetree of jawed vertebrates.
bioRxiv 2019.12.19.882829 (preprint)
doi: https://doi.org/10.1101/2019.12.19.882829
https://www.biorxiv.org/content/10.1101/2019.12.19.882829v1

6 thoughts on “Recalibrating clade origins, part 4

  1. Marjanovic often errors by not including extant taxa that are more primitive than extinct taxa that are older. This comes back to bite him several times, especially so when he relies on a single fossil tooth rather than a living animal he can hold. The LRT tests both living and extinct taxa to minimize taxon exclusion.

    As I’ve now commented on the first part, I did not conduct a phylogenetic analysis; I redated the phylogenetic tree found by Irisarri et al. (2017). That tree does have a few weak points, but that’s outside the scope of my preprint.

    That said, no way is Caluromys not a marsupial and Pteropus not a placental. No way. Feel free to add soft-tissue characters to your matrix if you can’t get the skeletal characters to work.

    Genomic studies deliver false positives

    You keep saying that, and you keep providing no evidence other than the LRT.

    So, again, we have a mismatch due to not testing all the mammals against all the mammals. That is what makes the LRT such a powerful tool that should be more widely used to avoid such old school mythology.

    If only your character sample was up to the task of resolving the phylogeny of your generally admirable taxon sample.

    • To your first point, Caluromys does have marsupial parts, but nests as the proximal outgroup to the Eutheria based on skeletal traits. One taxon has to be that one. Caluromys is that one. No one said Pteropus was ‘not a placental’. Don’t make up falsehoods to bolster your argument.

      To your second point, the LRT provides plenty of evidence that genomics delivers false positives. Genomic tests don’t always produce identical results, as you know. And genomic tests don’t always produce identical results to other phenomic tests, as you know. This is common knowledge now.

      To your third point, the character list has proven sufficient to lump and separate all adequately preserved taxa. That’s all you want it to do. The LRT provides a hypothesis or interrelationships. It is up to others to use similar taxon lists and their own characters to test these interrelations.

  2. One taxon has to be that one. Caluromys is that one.

    Why can’t it be Ambolestes or Juramaia or Murtoilestes or something like that?

    No one said Pteropus was ‘not a placental’.

    Sorry, I skimmed the text too quickly and overlooked this fact.

    Don’t make up falsehoods to bolster your argument.

    What is wrong with you that you immediately jump to the conclusion that people do things For Teh Evulz?!?

    Seriously. Do you honestly believe it’s more likely that someone lies than that someone just overlooked something?!?

    To your second point, the LRT provides plenty of evidence that genomics delivers false positives.

    Eh… I’m afraid this is circular.

    ‘Molecular data contradict the LRT, therefore they are wrong; molecular data are wrong, therefore it doesn’t matter that they contradict the LRT.’

    Genomic tests don’t always produce identical results, as you know. And genomic tests don’t always produce identical results to other phenomic tests, as you know. This is common knowledge now.

    Sure, except you’re committing the fallacy of false equivalence. There’s hardly any part of the vertebrate tree left where a consensus based on molecular data hasn’t emerged in the last 5 years, if not the last 20. The disagreements are getting smaller and smaller. The times when there were howlers like “Marsupionta”, or like mice and rats being found outside a “clade” of all other placentals, are decisively over.

    I complained that you never brought evidence for the inadequacy of molecular data other than the LRT. In response, you double down on the LRT, and add only that wrong trees have occasionally been inferred from molecular data. Is that all you can offer?

    the character list has proven sufficient to lump and separate all adequately preserved taxa. That’s all you want it to do.

    Uh… no.

    I don’t want to get a tree.

    I don’t want to get a resolved tree.

    I want to get as close as possible to the right tree. A poorly resolved tree (with polytomies) can be much closer to that than a fully resolved tree; that depends on the matrix, including the character sample. Like everything else, phylogenetics is not immune to garbage in, garbage out; I spent 10 years reducing the amount and proportion of garbage in 5 different matrices, and judging from the last time I saw your matrix, you still have work to do in that field.

    It is up to others to use similar taxon lists and their own characters to test these interrelations.

    Well, we’re working on it, but it’s so time-consuming to do this right that we can’t get it financed all at once. I started working on the me & Laurin (2019) paper in 2008 and was only able to submit it at the end of 2015. In one of those years I had to tutor highschool students for a living, meaning I couldn’t work on that paper full-time.

    • 1. Seriously. Do you honestly believe it’s more likely that someone lies than that someone just overlooked something?!? I cannot look into your soul. You produced a false statement for reasons you just now clarified. You can say it’s my fault. Or you can say, ‘you’re right. I produced a false statement. Sorry about that. Let’s move foreword.’

      2. It only takes one test to show genomics does not match phenomics. What should bother you is how quickly you accepted that the genomic clade that includes elephants + anteaters is basal to taxa that look like tree opossums. Or that if genomics and phenomics disagree, trust the one that does not indicate which clade have longer nasals than frontals, or the one that omits fossils.

      3. You wrote: “I don’t want to get a tree. I don’t want to get a resolved tree. I want to get as close as possible to the right tree.” The LRT is a virtually resolved tree (skull only and skull-less taxa cause problems) that produces a gradual accumulation of traits at every node. It’s size minimizes taxon exclusion. Whenever new data changes scores it is changed (the benefit of not being locked into paper media).

      • What do you mean by “basal to”? Atlantogenata comes out as the sister-group of Boreo(eu)theria – what sense does it make to claim one of two sister-groups is “basal to” the other?

        I cannot look into your soul.

        And yet you accused me of making things up to support a preconceived notion – of violating all kinds of professional ethics at enormous risk to my career, not to mention of trying to cheat my own desire to find out what really happened.

        Did you really believe that was at all a parsimonious hypothesis? Really?

        Or that if genomics and phenomics disagree, trust the one that does not indicate which clade have longer nasals than frontals, or the one that omits fossils.

        Longer nasals than frontals is just one character, and it correlates with lots of other things, so I’m not going to privilege it over the thousands of characters to be found in a genome.

        The LRT is a virtually resolved tree (skull only and skull-less taxa cause problems) that produces a gradual accumulation of traits at every node. It’s size minimizes taxon exclusion. Whenever new data changes scores it is changed (the benefit of not being locked into paper media).

        Good. Now merge the redundant characters, order the clinal characters, and add the hundreds of neglected characters, and it should become absolutely awesome. :-)

      • 1. re: more likely lying vs. overlooking something… I just correct errors of omission and commission, treating both the same.
        2. re: accused me… see #1
        3. re: Atlantogenata and Boreotheria… both are bogus, next?
        4. re: phenomics vs. genomics… gene studies cannot be checked for accuracy. You have to trust their results. False positives are possible. I prefer to have results you don’t have to trust because you can check them for accuracy.
        5. re: now merge characters, add characters, etc… not necessary. It may be flawed, but it’s working like it should, lumping and splitting far beyond its original intent. Now others can test their own character lists against this taxon list. Results, if valid on both sides, should match.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.