Friday, July 12, 2019

Phylogeny of the Lori analysis 2 - Topology

Now that we've covered my philosophy in designing the Lori analysis, let's look at some of the basic relationships it recovered.  The details are all laid out in the paper of course, and I'll write a post on each main clade as part of this series.

Something I'd like to clarify is that some commentary I've read has been to the effect of 'the support for lots of these clades is low, so be skeptical.'  That's true in the broad sense, but my point was that TWiG analyses over the past two decades also have low support for these clades, they just hardly ever report it.  You see e.g. Deinonychosauria again and again not because it's strongly supported in any TWiG analysis, but rather because scorings are reused and each analysis weakly supports Deinonychosauria based on the same weak character evidence.  So don't think the Lori analysis has unusually weak support for these clades, but rather that previous TWiG analyses don't report their support levels so you never knew how weak they were.

In any case, here's the full topology of taxa I consider strongly supported as maniraptoromorphs...

Strict consensus tree of Maniraptoromorpha after a posteriori exclusion of 43 taxa (see Positions of maniraptoromorphs pruned a posteriori in the supplementary info) (after Hartman et al., 2019).You'll probably want to click to enlarge...
Prior to Maniraptoriformes, we have a series of compsognathid-grade taxa including compsognathids themselves.  One surprise was finding Haplocheirus in Compsognathidae instead of Alvarezsauroidea.  As far as I know, this has only been previously suggested by Alifanov and Saveliev (2011:184) without proposed evidence, and based only on the initial description.  I used Choiniere's (2010) detailed thesis description to score the taxon, so incomplete scoring isn't an issue here.  I thought adding the new Xiyunykus and Bannykus might pull it toward alvarezsaurs, but nope, although scoring them in more detail after further description might influence things.  "Eight characters used by Choiniere et al. (2010) to place it in the latter clade were not included, but it also requires nine steps to constrain there in our analysis ... This suggests neither a compsognathid nor an alvarezsauroid identification is well supported and more study is needed."

The next 'big picture' concept is a fairly standard maniraptoriform topology of ornithomimosaurs branching first, then alvarezsaurs and therizinosaurs, then Pennaraptora.  What I found interesting about this area of the tree is that it's highly unstable.  As I wrote, "only four steps are required
to get a result similar to Sereno’s (1999) where alvarezsauroids are sister to ornithomimosaurs
and therizinosaurs sister to that pair. Similarly, while we recover a pairing of alvarezsauroids
and therizinosaurs to the exclusion of pennaraptorans, placing therizinosaurs closer
to the latter clade merely needs three additional steps. Positioning alvarezsauroids sister
to Pennaraptora or putting therizinosaurs just outside Maniraptoriformes are slightly
less parsimonious at six steps each..."  So Maniraptora is not the stable clade, Pennaraptora is.  For instance, moving therizinosaurs sister to oviraptorosaurs as used to be common in the late 90s-early 00s requires 13 steps while placing alvarezsaurs as paravians takes 11 steps.

Alternative topologies for major maniraptoriform clades in the Lori matrix showing the number of extra steps needed from the most parsimonious trees.
Similarly, there are a few taxa in this non-pennaraptoran maniraptoriform grade which have 'consensus positions' that are actually very poorly supported.  Pelecanimimus and Nqwebasaurus are ornithomimosaurs, right?  Well, Nqwebasaurus seems more likely to be an alvarezsaur as it "requires six steps to be an ornithomimosaur ... [and] all but two characters recovered by Choiniere et al. (2012) as supporting an ornithomimosaurian placement were included."  Pelecanimimus is more ambiguous, although it's most parsimoniously an alvarezsaur in my trees.  "Constraining it as an ornithomimosaur only requires two additional steps, however, where it emerges just above Shenzhousaurus as in Macdonald and Currie (2018). As only two of their characters supporting an ornithomimosaurian identification were not used by us, and only one from Brusatte et al. (2014), its true position is unclear pending a detailed osteology such as Perez-Moreno’s (2004) unreleased description."  The third controversial taxon of this kind is Fukuivenator, "emerging as the first branching alvarezsauroid, but moving to a basal therizinosauroid position with only two steps. A more stemward position seems more likely than a relationship with dromaeosaurids as suggested in its original description or Cau (2018), as it can be a coelurid with only four more steps, but takes seven steps to be sister to Pennaraptora and 11 steps to be paravian."  I'd now add Lingyuansaurus to this list, though it was published after submission time.  Recall it it a supposed therizinosaur, but moves to Ornithomimosauria with just three more steps.

Whatever Hesperornithoides is, it's clearly paravian, and paravian topology proved particularly labile in my analysis.  Your initial reaction might be 'hey, Xu et al. 2011 were right, and Archaeopteryx is a basal deinonychosaur."  Maybe.  But placing archaeopterygids (including 'anchiornithines') as avialans or sister to troodontids is only a step longer each.  Add to this scansoriopterygids, which are avialans in the most parsimonious trees but move to basal paravians in a single step, and troodontids, which are deinonychosaurs but move to Avialae in a single step, and the basic topology of Paraves is uncertain.  This isn't to say all alternatives are great though, as dromaeosaurids closer to birds than troodontids takes six more steps, and archaeopterygids basal to troodontids, dromaeosaurids and avialans takes 15 more steps. Oviraptorosaurian scansoriopterygids are 12 steps longer.  So we have a subset of plausible alternatives, and that's where the research should be focusing.  Stratigraphically basal deinonychosaurian archaeopterygids and basal avialan scansoriopterygids make sense, but who knows.

Alternative topologies for major paravian clades in the Lori matrix showing the number of extra steps needed from the most parsimonious trees.
The fifth major paravian clade is Unenlagiidae, traditionally paired with dromaeosaurids but here uniquely recovered as sister to Dromaeosauridae plus Troodontidae.   This Unenlagiidae includes halszkaraptorines as in Senter et al. (2012) and Cau (2018).  One interesting point is that the position of unenlagiids varies with the position of archaeopterygids- when archaeopterygids are sister to troodontids in trees one step longer, unenlagiids are sister to dromaeosaurids.  But when archaeopterygids are avialans in trees one step longer, so are unenlagiids as in Agnolin and Novas (2013).  Besides the standard set of taxa, we recovered the Gondwanan Pyroraptor and Ornithodesmus as unenlagiines as well as Dakotaraptor.  If the latter is true, is that a situation like Titanis and Alamosaurus where a southern taxon moved into North America?  New halszkaraptorines include the Early Cretaceous Ningyuansaurus and ISMD-VP09.

Finally, let's go over the basal avialan results.  I've always recovered Balaur as an avialan as in Cau et al. (2015) and it takes 8 steps to move to Dromaeosauridae.  Surprisingly, Hesperonychus groups with Balaur instead of microraptorians.  It only takes three steps to move to Microraptoria, but the stratigraphy is a better fit by Balaur.  "The branching order of Jehol non-ornithothoracine birds has been contentious, with our matrix supporting Sapeornis branching first, followed by jeholornithids then confuciusornithiforms. Jeholornithids branching first is only three steps longer, but Sapeornis branching last as in some recent analyses requires 12 more steps."  Between confuciusornithiforms and Ornithothoraces are Chongmingia, Yandangornis and Jinguofortis.  "... Jinguofortis joins Chongmingia in only three steps. Our analysis supports the latter’s position close to Ornithothoraces as in p2 of Wang et al.’s (2016) figure 7, whereas moving it to their p1 more stemward of Jeholornis and Sapeornis requires 11 more steps."

There are hundreds of other things to say about the topology, but next time I'm going to switch gears and discuss my first successful experience with peer review.  In the mean time, if any of you have questions about taxon placements or alternative topologies, feel free to ask.

References- Sereno, 1999. The evolution of dinosaurs. Science. 284(5423), 2137-2147.
DOI: 10.1126/science.284.5423.2137

Perez-Moreno, 2004. Pelecanimimus polyodon: Anatomía, sistemática y paleobiología de un
Ornithomimosauria (Dinosauria: Theropoda) de Las Hoyas (Cretácico Inferior; Cuenca,
España). PhD thesis, Universidad Autónoma de Madrid. 149 pp.

Choiniere, 2010. Anatomy and systematics of coelurosaurian theropods from the Late Jurassic of Xinjiang, China, with comments on forelimb evolution in Theropoda. PhD thesis, George Washington University. 994 pp.

Choiniere, Xu, Clark, Forster, Guo and Han, 2010. A basal alvarezsauroid theropod from the Early Late Jurassic of Xinjiang, China. Science. 327(5965), 571-574. DOI: 10.1126/science.1182143

Alifanov and Saveliev, 2011. Brain structure and neurobiology of alvarezsaurians (Dinosauria), exemplified by Ceratonykus oculatus (Parvicursoridae) from the Late Cretaceous of Mongolia. Paleontological Journal. 45(2), 183-190. DOI: 10.1134/S0031030111020031

Choiniere, Forster and De Klerk, 2012. New information on Nqwebasaurus thwazi,
a coelurosaurian theropod from the Early Cretaceous Kirkwood Formation in South Africa.
Journal of African Earth Sciences. 71-72, 1-17. DOI: 10.1016/j.jafrearsci.2012.05.005

Senter, Kirkland, DeBlieux, Madsen and Toth, 2012. New dromaeosaurids (Dinosauria:
Theropoda) from the Lower Cretaceous of Utah, and the evolution of the dromaeosaurid tail.
PLOS ONE. 7(5), e36790. DOI: 10.1371/journal.pone.0036790.

Agnolin and Novas, 2013. Avian ancestors: a review of the phylogenetic relationships of the
theropods Unenlagiidae, Microraptoria, Anchiornis and Scansoriopterygidae. Netherlands:
Springer. 96 pp. DOI: 10.1007/978-94-007-5637-3_1

Brusatte, Lloyd, Wang and Norell, 2014. Gradual assembly of avian body plan
culminated in rapid rates of evolution across the dinosaur-bird transition. Current Biology.
24(20), 2386-2392. DOI: 10.1016/j.cub.2014.08.034

Cau, Brougham and Naish, 2015. The phylogenetic affinities of the bizarre Late Cretaceous
Romanian theropod Balaur bondoc (Dinosauria, Maniraptora): dromaeosaurid or flightless
bird? PeerJ. 3:e1032. DOI: 10.7717/peerj.1032

Wang, Wang, Wang and Zhou, 2016. A new basal bird from China with implications for
morphological diversity in early birds. Scientific Reports. 6:19700. DOI: 10.1038/srep19700

Cau, 2018. The assembly of the avian body plan: a 160-million-year long process. Bollettino della
Società Paleontologica Italiana. 57(1),1-25. DOI: 10.4435/BSPI.2018.01

Macdonald and Currie, 2018. Description of a partial Dromiceiomimus (Dinosauria: Theropoda)
skeleton with comments on the validity of the genus. Canadian Journal of Earth Sciences. 56(2), 129-157. DOI: 10.1139/cjes-2018-0162

Hartman, Mortimer, Wahl, Lomax, Lippincott and Lovelace, 2019. A new paravian dinosaur from the Late Jurassic of North America supports a late acquisition of avian flight. PeerJ7:e7247. DOI: 10.7717/peerj.7247

32 comments:

  1. How stable is the grouping of Archaeopteryx with other archaeopterygids? This group has been alternatively known as Anchiornithidae so it's rather important to know how solidly Archaeopterygidae being the right name for it is.

    Sofar as Unenlagiidae goes, do you think it would be a good idea to redefine it to be excluded from Dromaeosauridae even in the traditional placement where it's considered a subfamily? Given this let's us always use the family name instead of switching all the time. Likewise, Unenlagiinae could become the name for the traditional grouping within the family, since Halszkaraptorinae is also included.

    ReplyDelete
    Replies
    1. It depends on your starting tree and what topology you're constraining. If Archaeopteryx is a deinonychosaur, forcing it to be basal to anchiornithines plus other deinonychosaurs takes 12 extra steps. If Archaeopteryx is an avialan, forcing it closer to Aves than anchiornithines takes 5 steps. I've recently scored each Archaeopteryx specimen seperately (except the Maxburg and chicken wing) and am in the process of scoring 8 Anchiornis specimens separately, so we'll have better data on this soon.

      Regarding Unenlagiidae, that's exactly what I did in the supplementary info- "Unenlagiidae- (Unenlagia comahuensis <- Archaeopteryx lithographica, Troodon formosus, Dromaeosaurus albertensis, Passer domesticus) NEW". I did the second thing as well- "Unenlagiinae - (Unenlagia comahuensis <- Halszkaraptor escuilliei, Microraptor zhaoianus, Velociraptor mongoliensis, Dromaeosaurus albertensis, Passer domesticus) NEW". Great minds think alike.

      Delete
  2. Correlated shifting of clades is also frequent in my analysis. Usually, some clades have a couple of favorite alternative placements, and move to one of these alternatives according to other clade movements. Usually, one options are: branches forming a paraphyletic grade A + others forming clade B, vs clade A + grade B.

    ReplyDelete
    Replies
    1. The situation described in the first two sentences has been very common in my analyses, too. I mention several cases somewhere in my huge paper.

      Delete
  3. Note that all unenlagiines have been found in deposits younger than the Cenomanian, and we'll have to wait for a newly described raptor dino from South America to confirm that Unenlagiinae and Halszkaraptorinae form a distinct clade to the exclusion of troodontids and all dromaeosaurids traditionally assigned to Microraptoria, Dromaeosaurinae, and Velociraptorinae.

    ReplyDelete
    Replies
    1. Ornithodesmus is Barremian, if my assignment is correct. Also, the Solnhofen Alcmonavis emerged there when I added it after submission. So we might get some earlier members, and both European interestingly enough.

      Delete
    2. Assuming that your initial placement of Ornithodesmus in Unenlagiinae is correct, then Ornithodesmidae would have priority over Unenlagiinae and Unenlagiidae, and Unenlagiinae could become a subfamily given Ornithodesmus' older age.

      Delete
    3. Also, the Solnhofen Alcmonavis emerged there when I added it after submission.

      Ha! That would make six origins of flight, two of them within Unenlagiidae/Ornithodesmidae! :-)

      Delete
    4. This comment has been removed by the author.

      Delete
    5. "Assuming that your initial placement of Ornithodesmus in Unenlagiinae is correct, then Ornithodesmidae would have priority over Unenlagiinae and Unenlagiidae, and Unenlagiinae could become a subfamily given Ornithodesmus' older age."

      Technically yes, but realistically Ornithodesmus is only weakly supported as going there, so I wouldn't recommend changing any family-level names until better evidence than the Lori analysis was presented.

      "That would make six origins of flight, two of them within Unenlagiidae/Ornithodesmidae!"

      Unless Alcmonavis was closer to Rahonavis than to flightless unenlagiines, or if unenlagiines were primitively volant and lost flight 1-4 times.

      Delete
    6. Speaking of the Alcmonavis paper, have you tested that ((Sapeornis+Jinguofortis)+Jeholornithidae) clade?

      Delete
  4. I can't Ornithodesmus in the cladogram listed in the blog post, even though Hartmann et al. mention it in the text. Any idea what happened?

    ReplyDelete
    Replies
    1. Ornithodesmus is among several taxa that had multiple equally parsimonious positions, so we pruned it a posteriori from the published cladogram. I list all of these taxa and their recovered range of positions in the supplementary info under "Positions of maniraptoromorphs pruned a posteriori." In the case of Ornithodesmus, it's an "Unenlagiine in the Pyroraptor+Unenlagia clade."

      Delete
  5. I would love some comment on the additional lengths to recover more traditional arrangements for eudromaeosaurs. Dromaeosaurus nested within asian "velociraptorine" taxa rather than Utahraptor and co. is one of the more bizarre placements in your tree for mine.

    ReplyDelete
    Replies
    1. An interesting thing about the dromaeosaurid part of the tree is that most suggested basic rearrangements happen in two steps or less. I found that Dromaeosaurinae is more inclusive than usual, and that Deinonychus and the Utah+Achillo pair make a third main clade. Rearranging these three clades into any combination takes 2 steps. Forcing Dromaeosaurus closer to Utah+Achillo than Deinonychus also takes 2 steps (IGM 100/22+23 follows).

      Delete
    2. oh interesting. Thank you for the answer.

      Delete
  6. “Gondwanan Pyroraptor and Ornithodesmus

    Did you mean “Laurasian”?

    Do you suspect that the basal placement of Lithornis is genuine (as Mayr has argued)?

    Do you trust oviraptorid Incisivosaurus or neoavian Paraortygoides?

    How many steps are needed for:

    1. Paravian Protarchaeopteryx?
    2. Avialan oviraptorosaurs?
    3. The traditional deinonychosaur/avialan dichotomy (with the latter including archaeopterygids & scansoriopterygids)?
    4. Palaeotis & Struthio as sister taxa?
    5. Anseriformes sister to Neoaves?
    6. Sauriurae?
    7. Galloanseres?
    8. Vegavids outside Neornithes?

    ReplyDelete
    Replies
    1. "Did you mean “Laurasian”?"

      That's a mistake for my trying to say Cretaceous Europe often resembles Gondwana more than Northern Asia or North America (E.g. spinosaurids, abelisaurs, common titanosaurs).

      "Do you suspect that the basal placement of Lithornis is genuine (as Mayr has argued)?"

      I think we'd need a matrix with more paleognath characters than mine and more non-crown taxa than published paleognath analyses.

      "Do you trust oviraptorid Incisivosaurus or neoavian Paraortygoides?"

      No and no, the former only takes 1 step to change while the matrix has a distinct lack of proposed neoavian or galliform characters.

      "1. Paravian Protarchaeopteryx?" 3, it's in the paperr.

      "2. Avialan oviraptorosaurs?" 12, it's also in the paper.

      "6. Sauriurae?" 51 to join Archaeopteryx and enantiornithines to the exclusion of Aves, 57 if you want Confuciusornis in there too.

      I'll have to test the others.

      Delete
    2. Some questions about more basal Coelurosauria, i.e. those more basal than the Ornitholestes+ clade shown in the cladogram.
      1) "several taxa usually considered members of [Compsognathidae] (e.g., Huaxiagnathus, Juravenator, Mirischia, Sinocalliopteryx) branch off more basally"
      I don't see Mirischia on the cladogram. Is that because it is more basal than Ornitholestes, or because it is difficult to pin down?

      2) "Megaraptorans and a clade of proceratosaurids and coelurids branch first after tyrannosauroids"
      Which was more basal?

      3) "megaraptorans or the coelurid-proceratosaurid group can be constrained to Tyrannosauroidea in only four steps"
      Is it that both groups simultaneously go to Tyrannosauroidea in only four steps? Or each group separately? If separately, how many steps for both megaraptorans and coelurids+proceratosaurids to end up in Tyrannosauroidea?

      4) Is the Proceratosauridae you recovered limited to the usual suspects (Proceratosaurus, Guanlong, Sinotyrannus, Kileskus), or does it include others (e.g. Juratyrant, Stokesosaurus), as in Loewen et al. (2013)?

      5) Where did you find Bicentenaria, Zuolong, and Chilesaurus?

      Lastly, The cladogram only shows Ornitholestes+, but your analysis was much larger. Did you exclude the rest just for the sake of space? or because you're worried readers won't heed your warning that the matrix is not well-suited for assessing the relationships of non-Maniraptoromorphs?

      Delete
    3. First, the mpt is in the tnt file, so anyone can view it. I'll post about it here at some point.

      "I don't see Mirischia on the cladogram. Is that because it is more basal than Ornitholestes, or because it is difficult to pin down?"

      It's a proceratosaurid/coelurid.

      ""Megaraptorans and a clade of proceratosaurids and coelurids branch first after tyrannosauroids"
      Which was more basal?"

      Megaraptorans.

      "Is it that both groups simultaneously go to Tyrannosauroidea in only four steps? Or each group separately? If separately, how many steps for both megaraptorans and coelurids+proceratosaurids to end up in Tyrannosauroidea?"

      Separately in 4 steps each. Have yet to check if both are forced there at once.

      "Is the Proceratosauridae you recovered limited to the usual suspects (Proceratosaurus, Guanlong, Sinotyrannus, Kileskus), or does it include others (e.g. Juratyrant, Stokesosaurus), as in Loewen et al. (2013)?"

      Proceratosaurids and coelurids mix together. Sinotyrannus is sister to Yutyrannus in Tyrannosauroidea, while Juratyrant is sister to Eotyrannus in Tyrannosauroidea. Stokesosaurus was excluded as it was too unstable.

      "5) Where did you find Bicentenaria, Zuolong, and Chilesaurus?"

      In the saved tree, Bicentenaria is a eustreptospondylid megalosauroid and Zuolong is the most basal megaraptoran. I can't remember if they can fall out in other places though. Chilesaurus was not included since I have no confidence its an averostran, but came out as an alvarezsauroid in an earlier version. I still have it scored, so I could check.

      "The cladogram only shows Ornitholestes+, but your analysis was much larger. Did you exclude the rest just for the sake of space? or because you're worried readers won't heed your warning that the matrix is not well-suited for assessing the relationships of non-Maniraptoromorphs?"

      Closer to the latter. They were included to help attract potential coelurosaur mimics like Afromimus and to provide a decently extensive outgroup to polarize Coelurus', Tanycdolagreus' and Ornitholestes' character states, to help test Rauhut's idea Lori could be a juvenile of those. But the topology itself is not based on good assumptions, so it's not something I wanted confused for a useful result.

      Delete
  7. What happens when “Protoavis” is added to the matrix (assuming Chatterjee’s description is accurate)?

    ReplyDelete
    Replies
    1. It isn't accurate. The neck and probably the head belong to a drepanosaurid; the ilium is pieced together from tiny fragments, and the space between them filled with glue; and so on.

      Delete
    2. I would need to score Protoavis. Which sounds like a fun project and post, but merely out of historical interest since David's correct that the Protoavis animal hypothesized by Chatterjee isn't a thing.

      Delete
    3. A fun project? Sounds like a world of pain.

      Delete
  8. Sorry for double-commenting, but how many steps are needed for arctometatarsalian Avimimus or Xu et al.’s (2010) alternative pennaraptoran topology?

    ReplyDelete
    Replies
    1. Haven't checked Avimimus, and which Xu et al. topology do you mean? The Xiaotingia one?

      Delete
  9. Apologies for indulging a personal tic and beating a dead language to death, as I know your specialty is phylogeny and not alpha taxonomy or Clessical languages, but the statement, from the *Hesperornithoides* description, that "[t]he trivial epithet honors the Miessler family" is not quite accurate.

    That may have been the intent, but in actuality the form *miessleri* can only honor a single male individual of that family.

    ReplyDelete
    Replies
    1. I brought this up in review. (Hey Scott, I told you someone would say this!...)

      Delete
    2. Aaaaarrrrrrrrgh. I overlooked this.

      This is one of the many, many cases where the ICZN (Articles 31–33) is unclear on whether this is a correct or an incorrect original spelling.

      Seriously, how hard can it be to keep us out of that mess? English spelling distinguishes 's from s', too!

      Delete
    3. I've seen it argued that since "the Miessler family" is a singular unit, the "-i" suffix is technically acceptable. Much like how, say, Vegavis iaai uses the "-i" suffix, even though "the Instituto Antártico Argentino (IAA) expedition that collected the specimen" was made up of multiple people. Whereas had the etymology been "for Howard and Helen Miessler," the "-orum" suffix would be mandatory. At least, that's what Peter Dodson claimed in defense of the original spelling of Avaceratops lammersi, in his book The Horned Dinosaurs. The current ceratopsid literature does call it "lammersi."

      Delete
  10. This comment has been removed by the author.

    ReplyDelete
  11. Off a skim of the paper and post it's never addressed how solid the grouping of Unenlagiinae and Halskaraptorinae is - how easy is it to move Halskaraptorinae to any number of other positions in Deinonychosauria?

    ReplyDelete