I would be grateful if somebody could help me out here as Bioinformatics is not my strong card.
Regarding the recent Nature publication
New deep-sea species of Xenoturbella and the position of Xenacoelomorpha
I query the authors’ explanation; to wit…
The sister group relationship between Nephrozoa and
Xenacoelomorpha supported by our phylogenomic analyses implies that the last common ancestor of bilaterians was probably a benthic, ciliated acoelomate worm with a single opening into an epithelial gut, and that excretory organs, coelomic cavities, and nerve cords evolved after xenacoelomorphs separated from the stem lineage of Nephrozoa.
My problem arises with their placement of Ctenophora on their own phylogenetic tree as the “more primitive out-group” (for lack of better words on the spur of a rushed moment). Myself, I always considered Ctenophora as bilateral – in this case more primitively bilateral which IMHO should root the bilateran tree… which of course begs more than one question upon rereading their analysis.
Forget Ctenophores – what about Cnidarians!? Some taxonomists argue that Cnidarians are descendents of ancient bilateral coelomates and not the other way around. Biologists have known since the 1920s that Cnideria had a directive axis which gave them right and left-hand sides. Volker Schmidt goes on to argue that non-radially organized hydrozoan larvae have an anterior concentration of sensory and ganglionic nerve elements, suggesting that a fundamental genetic toolkit for the establishment of bilateral and polarized anatomies was already present before the Cnidaria-Bilateria divergence. Volker Schmidt goes so far as to suggest that diploblastic status of adult Cniderians is derived and that true mesoderm can be even be detected during Cniderian embryogenesis. OK – I concede that last argument is particularly contentious… but you get my drift.
I am partial to the notion the UrBilateran that subsequently gave rise to “Protostomes” & Deuterostomes and was itself coelomate with possessed a dorsal nerve chord. Any subsequent acoelomy and pseudocoelomy was derived… ditto ventral nerve chords. But hey… now I am being really contentious!
Apologies for not noticing this earlier (as a “pending” post). It’s best to put a comment in the moderation thread to alert us about a pending new topic.
Maybe someone can write Tom Meuller and tell him what happened?
Neil Rickert,
I alerted Tom Mueller and edited the publication date so the OP is on the front page.
Tom should correct the typo for “latest” in the post title.
It seems to me that there are two issues; (1) what is the tree, and (2) where on it does bilaterality evolve.
The placement of Ctenophora is an issue of considerable and fascinating debate:
http://www.pnas.org/content/112/50/15402.abstract
And in reply:
http://www.pnas.org/content/113/8/E946.extract
Joe Felsenstein,
Hi Joe,
I am cyber-UNsavvy… I somebody could tell me how, I would gladly oblige!
best
You most likely need an admin to give you editing privileges.
Hello all… I had forgotten I had even posted the following and have since become even more confused.
Here are two relevant but contradictory links making me feel like some proverbial motionless starving donkey transfixed between two equidistant bales of hay.
http://mprous.blogspot.ca/2016/02/xenoturbella-and-xenacoelomorpha.html
https://medium.com/@caseywdunn/who-is-our-most-distant-animal-relative-4309b9fe4e48#.rq68u89bx
It appears there are two contending schools of thought polarizing about either the Pisani et al group or the Dunn et al group.
Here is my understanding on what bilateral is supposed to mean:
Humans are bilateral because in addition to top and bottom, they have dorsal and ventral sides and mirror image left and right.
Echinoderms are bilateral because as larvae they were no different than humans and as adults they are pentamerously bilateral (not pentamerously radial as stated in texts)
Cnidaria are bilateral because they their planula larvae are as bilateral as human adults and as embryos Cnidaria have a directive axis of bilateral symmetry.
Ctenophores are technically NOT bilateral because even though they posses top and bottom, and mirror image left and right Ctenophores do not have distinguishable dorsal and ventral sides which instead are again mirror images. I am not clear if their internal anatomy can indeed bestow true ventral vs dorsal distinctions, but I suspect not.
Introductory textbooks are hopelessly confused on all this.
Meanwhile, Kevin J. Peterson’s work with miRNA is exciting. It appears miRNA may be refractory to Long Branch Artifacts if I understand any of this correctly.
That is why I am intrigued by the suggestion that Xenoturbella’s “primitive status” may in fact be derived.
http://www.mpg.de/1164493/deuterostomia_phylogeny_xenacoelomorpha?filter_order=LT&research_topic=BM-EB%2CBM-EVB%2CBM-G
I am convinced the primitive UrBilateran was probably pretty complex but rapidly split into different lineages in response to what Hejnol & Martindale deem the shift of blastopore formation from the animal pole to the vegetal pole creating developmental problems that required quick resolution. I resorting to pure speculation here, but perhaps the diffuse nerve net of the UrTriploidBilateran ancestor needed to concentrate ventrally or dorsally due to developmental constraints imposed by the blastopore axis shift,an either/or commitment that bifurcated the Ecydysozoa/Lophotrochozoa vs Deuterostome lineages.
I still remain confused regarding the definition of Bilateral. Aren’t the Planulae of Cnidarians bilateral? So why are not Cnidaria considered bilateral no differently than Echinoderms?
Meanwhile – I become more and more perplexed by Xenoturbella. Does Xenoturbella embryology indeed recapitulate Lophotrocophora ontogeny? If so – wow!
http://rspb.royalsocietypublishing.org/content/266/1421/835
… and back to the drawing board we go!!!
At some point I throw my hands up in despair and wonder if it is possible to make any sense of the disparate and contradictory lines of evidence! The inclusion of some data to the exclusion of other, to support one interpretation instead of another, begins to look pretty arbitrary to my jaundiced eye.
Thanks in advance to any and all for any help on the matter.
Because they are not like humans. I blame anthropocentrism!
TomMueller,
I fixed the typo. A trace of it remains in the URL for the topic, which can’t be changed because that would mess up any links to the thread.
Neil Rickert,
TomMueller,
My apologies for messing up the format of the last comment…
Paleontologist Marcus Ross had this to say about the Protosome & Deuterostome ancestor. It may not help resolve your problem, but make it worse, but at least you’ll know you’re not alone in your difficulties. Ross says the creature would look like something from the X-files. It’s one of my favorite papers over the last decade:
http://digitalcommons.liberty.edu/cgi/viewcontent.cgi?article=1079&context=bio_chem_fac_pubs
stcordova,
Of course, Marcus Ross is a young-earth creationist. How a paleontologist could be a YEC is quite beyond me, and it’s also beyond me why we should credit anything he says, since he thinks his main source of data is just a product of a single contemporaneous biota. I would however be interesting to see not just you quoting his opinion but you, in your own words, justifying that opinion and explaining just what it means and how it relates to the current discussion.
It’s a beautiful con game you guys have going on here.
Quote a creationist source and it can’t be trusted. Quote a non-creationist source and you’re quote-mining.
Ctenophora – Their phylogenetic position is far from clear.
– The Tree of Life
Yes, that would certainly be reprehensible if anything like it ever actually happened.
Mung,
So, are you a YEC? Do think the YEC claims ought to be considered scientifically respectable? Do you think a flat-earthe’s claims about geography deserve serious consideration?
Are you even aware of what you did there?
Tom was frustrated, and I was just pointing out maybe he’s trying to solve an unsolvable problem — like square circles in Euclidian geometry.
I thought Marcus Ross and Paul Nelson showed legitimate mechanical reasons that are barriers to the UrBilatarian that Tom was talking about.
The article showed how others have tried and failed to characterize UrBilatarian, maybe it will never be characterized.
No I am not a YEC. I think YEC is silly. To the extent that flat-earthism is based on observation and not on something written in the Bible YEC is worse than flat-earthism.
Possibly not. Can you tell me?
So why shouldn’t we discount what a YEC has to say about paleontology? Shouldn’t we at least ask him to justify his claims?
Not in any way a response to any of my questions. The article you referenced (itself unpublished) offers no mechanical reasons at all. And in fact there has been vast progress in bilaterian phylogeny in the past few years. We have confidence in the existence of, composition of, and relationships among Deuterostomia, Lophotrochozoa, and Ecdysozoa. While there continue to be enigmatic relationships, there is no reason to believe they will be forever intractable. Ross’s belief to the contrary seems due to an a priori belief in separate creation of “kinds”, whatever those are.
It’s hard to discuss any subject of this sort with a person whose first principle is that the universe is 6000 years old. It’s even harder because you don’t want to talk about that.
Gentlemen, this may not be my blogsite – but this happens to be my blog. Could I ask you to take your argumenta ad hominem elsewhere?
Joe Felsenstein laconically remarks:
It seems to me that there are two issues; (1) what is the tree, and (2) where on it does bilaterality evolve.
Exactly! Herculean attempts to identify modern ambassadors of ancient lineages that indeed root the metazoan ToL is beginning to appear a “fool’s errand”!
Meanwhile, it appears that obeisance to inconsistent older convention is taking precedent to logic when defining terms such as “bilateral”…
I compare this exercise to the employ of logic to determine the original order of a deck of cards after they have been shuffled. At some point the cards are shuffled so much it becomes impossible to complete the task.
My indignation at the pervasive intellectual incest of standard textbook boilerplate got me started on all this when I realized the following:
Some coelomate Locotrophozoan ancestor gave rise to pseudocoelomates (today’s Rotifers). Some coelomate Ecdysozoan ancestor also gave rise to peudocoelomates (today’s Nematodes). Some coelomate Locotrophozoan ancestor also gave rise to acoelomates (today’s Flatworms).
I remains still within the realm of possibility that some coelomate Deuterostome ancestor also gave rise to acoelomates (Xenoturbellida and Acoelomorpha)…
In other words, all modern acoelomates and pseudocoleomates are probably derivative. Meanwhile, it is becoming more and more likely that the original Ur-Triplobalstic-Bilateran may indeed be ancestral to Cnidaria and Porifera.
If we can accept that Bilterality was preceded by an intermediate form of “pseudo-bilaterality” (as in Ctenophores”) then much of what I read becomes over-enthusiastic hubris.
Perhaps Jordi Garcia-Fernàndez nailed it when he claimed
“…in fact, the simplest bilaterian metazoans, having a single or only a few Hox-like genes and several NK genes. This would imply that no basal non-bilaterian animals currently exist. An intriguing exception might be the placozoans…”
http://www.nature.com/nrg/journal/v6/n12/full/nrg1723.html
John Harshman,
Hi John.
You state that
And in fact there has been vast progress in bilaterian phylogeny in the past few years
I would be grateful if you could direct me to any reference you think may clarify my obvious state of confusion.
FTR – I remain in your debt for earlier intercession of yours that have corrected fuzzy thinking on my part.
best & grateful regards
I’m not talking about Xenoturbella but about the major branches of Bilateria that are mentioned in the comment you quote, right after the sentence you quote.
Looking over Casey Dunn’s précis, it would appear that data can be “massaged” – cross that – “blatantly manipulated” i.e strategic inclusion or exclusion of data can lend support to one contention or another.
Of course, I must be over-simplifying this all.
Forgive this shameless bump, but I would be grateful for any clarification on the matter.
best
TomMueller,
It would help if you asked more specific questions than “What’s all this then?”. And it would help if you explained how you arrived at the understanding you do have, more than “looking over Casey Dunn’s precis”.
John Harshman,
To answer John VeryHarshMan’s criticism
It would help if you asked more specific questions than “What’s all this then?”. And it would help if you explained how you arrived at the understanding you do have, more than “looking over Casey Dunn’s precis”.
OK – have it your way! Here is Casey Dunn’s take on the analysis by the Pisani group:
…they used different models of molecular evolution, they changed outgroup species sampling, and they changed gene sampling… All analyses under consideration include closely related outgroups, so the variation under consideration is whether more distant outgroups are included or removed… I am highly suspect of being asked to reject a hypothesis that is recovered with broader taxon sampling because reduced taxon sampling supports a different hypothesis… Much of the disagreement regarding the placements of sponges and ctenophores now comes down to how to interpret sensitivity to outgroup sampling, and how this may interact with model selection.
Any high school appreciates that so-called “outgroups” cannot be presumed but rather need to be deduced. The rooting of trees using outgroups is problematic if our assumptions that they are actually outgroups is incorrect.
Which brings me back to my original frustrated contention that seems to have so irked John Harshman (damn he seems to be a grumpy cuss, I am beginning to feel sorry for him)
Looking over Casey Dunn’s précis, it would appear that data can indeed be “massaged” – cross that – “blatantly manipulated” i.e strategic inclusion or exclusion of data can lend support to one contention or another.
I will not repeat in detail all Dunn provides by way of support for his conclusions (this post is already too long and I do not want to insult anybody’s intelligence) nor will I repeat in detail all the contradictions of Casey Dunn by Marko Prous cited above ( I assume everyone present possesses reasonable reading comprehension)
sooo…
My original question remains standing: Is it in principle even possible to root the metazoan tree, or has that for all intents and purposes become a “bridge too far”? … and would any answer be suspect given underlying premises (by one camp vs the other) appear to be presumed and not deduced? (you will need to go right back to the top John – for the sake of brevity I did not repeat myself, mind you as I remember it, when I in fact do repeat myself, somehow that also seems to provoke you as well – just no pleasing you is there, but we digress)
Considering all the contradictory and disparate line of evidence, including as just one “for example” the suggestion that Xenoturbella embryology may even recapitulate Lophotrocophora ontogeny, my frustration remains quite unabated and not for reasons John AllTooHarshMan would suggest… I am overwhelmed by too much contradictory detail and my line of questioning has in fact, if anything, been bogged down by wallowing too much “on specifics”.
But hey, I am grateful for any bump, from any corner.
At least John is not accusing me of “hyperbole or belief in some weird things” this time around… hopefully I will not now be accused of Sesquipedalian Loquaciousness.
Hey John, I hope you and I meet sometime, I really would like to buy you a drink.
In any case, I confess this is not my field of expertise, so any elucidation from any and all gratefully appreciated. I thank you in advance.
True. This is not, however, the point anyone is making. All the outgroups in question are agreed to be actual outgroups by everyone. The question is whether it’s better to use just the sister group of the ingroup or to use successively more distant outgroups. I think Dunn is wrong about this: the best possible outgroup choice is a densely sampled sister group of the ingroup. Adding more distant taxa actually can (though won’t necessarily) reduce your chance of a correct topology or rooting.
Well of course it can. But is anyone doing that? Is there any accusation in the piece that anyone is doing that? So far I don’t see it.
The answers are yes and no. It’s difficult, given the ages of the events. And there are tests that should allow one to decide that one topology is superior to another, given the data, if indeed one is. Note, for example, that Dunn addresses the basic methodological conflict by showing that for at least one data set, choice of outgroup does not change the topology.
Nevertheless, specifics are the only place in which clarity might be found. One can only look at the data and analyses, think of ways to test their assumptions, and perform those tests. And of course, find more data.
John Harshman,
Hello John
Thank you for your considered reply. I bear you no grudge and I hope you bear me none as well. The fact is that you have been a great help to me in the past as now and for that I thank you.
I appreciate what you are saying.
You state: All the outgroups in question are agreed to be actual outgroups by everyone. The question is whether it’s better to use just the sister group of the ingroup or to use successively more distant outgroups.
Well I wonder if in fact we can all be certain whether we all have that story right, and whether some question-begging is going on. Identification of outgroups can be problematic. Our job would be easier, for example if Choanoflagellates and Ctenophores possessed miRNA. ITMT I have always wondered if Choanoflagellates may be reiterating the single-celled Yeast story, where so-called “primitive” single cells actually were descendants of so-called “more complex” “haphaenated” (nyuck nyuck) multi-cellular ancestors.
Mind you, this following paper seems (emphasis on seems) to have confirmed the status of Choanoflagellates as sister group.
http://www.pnas.org/content/105/43/16641.full
Along these lines, I think was you who alerted me to the possibility that modern cartilaginous fishes (at least some of them) may be descendants of boney fishes. (…like I said – I do remain grateful to your patient intercessions)
I find your next suggestion very intriguing:
The question is whether it’s better to use just the sister group of the ingroup or to use successively more distant outgroups. I think Dunn is wrong about this: the best possible outgroup choice is a densely sampled sister group of the ingroup. Adding more distant taxa actually can (though won’t necessarily) reduce your chance of a correct topology or rooting.
That is exactly the point that Marko Prous (cited above was making)
The genome content tree that Ryan et al. (2013) recovered, contained several phylogenetic relationships that are highly suspect (e.g. non-monophyly of Annelida and a group containing a mollusc, an annelid and a chordate Branchiostoma). Remarkably, when Pisani et al. (2015) used a less biased model of gene content evolution, which does not underestimate gene losses as much as did the model used by Ryan et al. (2013), they recovered an animal tree of life that fully agreed with independent phylogenomic analyses based gene sequences. There was not a single nonsense clade left. And in that tree, Porifera replaced Ctenophora as the sister group to other animals.
Marko Prous would seem to confirm your version of events, making Casey Dunn’s thesis perhaps somewhat questionable.
If you recall, you and I touched on this topic on Larry’s forum. Larry lamented (more than once) how different genes tell different stories and as far as he is concerned there is no unique ToL. Simon Gunkel & Georgi Marinov emphatically disagreed, at least for eukaryotes.
http://sandwalk.blogspot.ca/2015/05/ford-doolittle-talks-about-tree-of-life.html
http://sandwalk.blogspot.ca/2015/05/molecular-evolution-exam-april-2015.html
According to Georgi:
… such lineages begin to resolve better, and once you have many genomes, the picture improves considerably
My reading of Koonin would made me tend to agree with Georgi in principle, but I now am beginning to wonder. Sure there has gotta be one unique ToL, that part seems clear. What remains (at least to me) unclear is whether we really will be in any position to correctly deduce it.
In any case, your response prompted some google-whacking. I did not appreciate how failure to identify “closest outgroups” in taxon selection could quickly lead to Long Branch Attraction problems. Forgive me if I am beginning to sound repetitive, but at some point, I wonder if we are beginning to chase our tail. Yeah we need many more genomes; but no – we still require “closest outgroups”; but which are those again…?
As Casey Dunn also explains in his excellent review, gene sampling, and analysis method used are also important and not just taxon sampling as you mentioned.
http://www.cell.com/trends/ecology-evolution/abstract/S0169-5347(15)00062-2
My frustration is that I do not have the background knowledge to assess the merits and weaknesses of different approaches to these three parameters, for example how to assess the pros and cons of a CAT vs WGT model… as overviewed here:
http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1000602
In any case your point was well taken – I should have been more specific, even though the intent of my preceding penultimate post was a brief “bump”.
In any case, Casey Dunn make an excellent case that much more work needs to be done in order to reveal the “hidden biology” of sponges and ctenophores in order to better elucidate relationships between taxa. Something along these lines was recently attempted by Gáspár Jékely with respect to the evolution of nervous systems.
http://evodevojournal.biomedcentral.com/articles/10.1186/2041-9139-6-1
Jékely’s paper also supports your contention that the inclusion of more distant outgroups can create false topologies by inevitable generation of LBAs. However, this paper still leaves open the possibility that Casey Dunn’s ctenophore as metazoan sister group may still be salvageable.
I love this diagram
https://static-content.springer.com/image/art%3A10.1186%2F2041-9139-6-1/MediaObjects/13227_2014_126_Fig2_HTML.jpg
I am guessing that Jékely’s paper’s may have prompted Casey Dunn’s review given how Dunn waxes eloquent on Poriferan “action potential” among other hitherto “hidden biology”.
best and grateful regards
Tom, I find it hard to reply to you because your comments seem unfocused and scattershot. We may indeed wonder whether choanoflagellates are truly an outgroup to metazoans. The way to find that is to add another outgroup that we are sure of. There happen to be quite a few studies that incorporate various metazoans, at least one choanoflagellate, and at least one other outgroup; I am aware of none that has contradicted the standard story. (Note that no topology necessarily implies either that they are primitively unicellular or secondarily so; that’s a separate question.) My point, however, wasn’t that we know choanoflagellates are an outgroup. It was that none of the papers you have cited takes issue with that claim, while you had implied otherwise.
Of course many parameters are important in phylogenetic analysis; I didn’t intend to imply otherwise. I was just concentrating on the parameters that were relevant to the immediate point.
I think one of your problems here is confusing ancestral character states with phylogenetic relationships. These are quite separate issues, though one can help in estimating the other. Still, groups are defined based on the latter, not the former. A species can be a metazoan without being multicellular, a protostome without having its mouth homologous to the blastopore, and a tetrapod without four legs. (In principle; we have known examples only of the last two.) For clarity, you need to separate these two matters in your comments.
Hello John – I return from a skiing holiday only to discover more cantankerous condescension on your part!
John Harshman
We may indeed wonder whether choanoflagellates are truly an outgroup to metazoans. .
Exactly!!!!, given historical distraction since 1866 of confusing presumed ancestral character states with phylogenetic relationships, i.e the similarity of sponge choanocytes to unicellular choanoflagellates.
John Harshman
The way to find that is to add another outgroup that we are sure of. .
Yeah, except you are contradicting what you just said earlier.
John Harshman
(Note that no topology necessarily implies either that they are primitively unicellular or secondarily so; that’s a separate question.) .
Yeah, I know! Exactly my earlier point, explaining why I referenced the Casey Dunn paper where he specifically said:
It is generally [incorrectly] considered that these collar cells are a synapomorphy of choanoflagellates and animals…
John Harshman
My point, however, wasn’t that we know choanoflagellates are an outgroup. It was that none of the papers you have cited takes issue with that claim, while you had implied otherwise.
Again you are being deliberately obtuse! Again I quote from Dunn which I cited earlier for good reason:
…Choanoflagellates are of great interest as the outgroup to animals, regardless of the homology of collar cells.
Not to mention I also cited this paper: http://www.pnas.org/content/105/43/16641.full
… so no – I made no such implication and I resent you putting words into my mouth.
John Harshman
I think one of your problems here is confusing ancestral character states with phylogenetic relationships.
Godamit – now you are being deliberately obtuse or you have a serious deficit in reading comprehension! The whole reason I started the thread to begin with was precisely because I recognized that was not the case. I was hoping that others with more expertise than I would be able to resolve some apparent (to my amateur eyes) contradictions in interpreting phylogenomic data. Read the title fer cryin out loud!
John Harshman
A species can be a metazoan without being multicellular…
How were you supposed to be contradicting me again?
John Harshman
…a protostome without having its mouth homologous to the blastopore, and a tetrapod without four legs. (In principle; we have known examples only of the last two.)
Godamit again – maybe it is not your reading comprehension at issue here but a deficit in memory!
On more than one occasion I have said the very same thing, especially in deference to Hejnol & Martindale’ excellent analyses on the subject and more than one have suggested that boiler plate textbook distinctions between protostomy vs. deuterostomy can be tossed out the window when I suggested (again on frequent occasion) that the position of the ventral vs dorsal nerve cord was probably the last criterion left standing in distinguishing deuterostomes from their lophotrochozoa & ecdysozoa cousins. Just read my previous remarks!
Before I weigh in to contradict your other non sequitors, perhaps you would care to reread what I actually said and what was actually said in the papers I cited. Scratch that – don’t bother. Our exchanges really have become futile!
I have gone out of my way to be conciliatory and you have stubbornly remained steadfastly condescending. I was hoping your expertise in phylogenomics could have been helpful but you continue to stoop to personal attack-mode.
For the benefit of bystanders: John Harshman is a PhD evolutionary biologist. I am an aging and hopelessly non-current High School Biology teacher. John is a professional – I am less than an amateur.
For some reason that I cannot fathom, John seems to take delight in drive-by snide sniping to my admittedly naïve and non-expert lines of inquiry and by putting me in my place.
For example, on the CRISPR goes retro thread:
John Harshman
Either you have a tendency to hyperbole or you believe some weird things.
When I explained my regrettable typo in polite terms:
Hi John, you are correct, an inversion is not the same thing as an insertion. My bad…Although technically speaking, regardless of whether one invokes the replicative or the conservative model of transposition, both amount to the same thing.
Again John Harshman persisted with drive by snide sniping:
John Harshman
They do not. What?
Of course, it rapidly became clear that I was correct, John dropped the matter without a word of acknowledgement only to lie in wait to snipe yet another time.
This puerile behavior is a repeat performance from other boards.
Of course, I am no PhD and my naïveté does on occasion require correction (again please refer to the title of this thread).
Perhaps I am also guilty of occasional hyperbole – no less than Tom Cech apparently, but the insulting charge of my belief in “weird things” clearly does not stand. Let’s say I prefer to be placed in the company of Tom Cech than John Harshman.
I thank Allan & David & Rumraket and others for their patience and their indulgence. Gentlemen, I remain in your debt. John, I really hope our paths do cross one day so I can buy you that drink. I really am curious and would like to see what makes you tick.
Meanwhile, I would like to ask others for advice on how best to deal with this bit of unpleasantness, given this forum seems decidedly more civil with higher expectations than others.
For example, exactly how does the “ignore” feature work?
John Harshman,
Really!?
@ Tom
You may not have read the site rules. The aim is to attack ideas, not people.
Alan Fox,
Hi Allan,
You wrote
“You may not have read the site rules. The aim is to attack ideas, not people.”
I apologize if I have crossed the line. If so, please delete my post.
Best regards,
tm
TomMueller,
I responded here.
Tom, if you don’t want me to talk to you I’ll be happy not to.
John – I will be frank.
Any exchange between the two of us will be unfair given you have far more to offer me than the reverse. I appreciate that.
However, I would much prefer we rebooted our exchanges such that they could continue in courteous fashion.
I have in past benefited from your intercessions and would be eager to learn from you in future.
I just do not want to continue the futile sandwalk acrimony here. That’s all.
Can we bury the hatchet, let bygones be bygones and move on?
I would be delighted if you were OK with that…
best regards
I don’t recall any acrimony. And I have never attempted to be rude to you. But I am often blunt.
Hi again John –
Sure, OK, whatever: Water under the bridge, let’s move on.
John Harshman,
Here is a quick and hopefully focused query;
Earlier you said,
The question is whether it’s better to use just the sister group of the ingroup or to use successively more distant outgroups. I think Dunn is wrong about this: the best possible outgroup choice is a densely sampled sister group of the ingroup. Adding more distant taxa actually can (though won’t necessarily) reduce your chance of a correct topology or rooting.
Would it be possible (according to what you just said) that choanoflagellates are thereby disqualified for consideration as an outgroup when attempting to sort out Xenoturbella’s phylogenomic relationship to even more recently diverged lineages?
Is it possible that no one may be in any position to presume the immediate sister group of Xenoturbella leaving no choice but to end up begging a very frustrating question?
Again, I appreciate that I am again asking very naïve questions.
Not disqualified, just perhaps not the best. There are however reasons why it still might be the best. If you don’t know what the relevant sister group is, i.e. if you aren’t sure what group Xenoturbella belongs to, best to be safe and choose one you know is an outgroup, even if it isn’t the immediate outgroup. The other reason is if the immediate outgroup has a low number of species or all the species are recently diverged, either of which would make dense sampling of a useful sort impossible. Given the ambiguity at the base of Metazoa, it would probably be wise to use choanoflagellates. Lots of them.
Hi John,
My apologies for missing your reply.
I think I understand our answer. Where I am having difficulty is reconciling what you just wrote with your earlier reply and I quote:
JH: I think Dunn is wrong about this: the best possible outgroup choice is a densely sampled sister group of the ingroup. [Not an option for reasons you just explained] Adding more distant taxa actually can (though won’t necessarily) reduce your chance of a correct topology or rooting.[which to my reading sounds a lot like what Casey Dunn is saying]
At first glance, it appears to me you are contradicting yourself and end up reaffirming Dunn’s contradiction of Pisani.
I must be missing something. Perhaps the selective criteria of including and excluding different outgroups is what’s at issue here – such that “nonsense clades” are not generated.
I would be grateful for any clarification of my obvious confusion.
Thanks in advance.
Hi again John…
Forgive this shameless bump – but one question remains unanswered. Any other takers, if John is unavailable?
thanx in advance
TomMueller,
First, I don’t understand why you think I have contradicted myself. What looks contradictory to you? The idea is simple: the best way to root a tree (and to get a valid topology) is to sample taxa such that every internal branch is short but not too short. This means that there should be short branches in the outgroup as well as the ingroup. The best way to ensure that the branches in both groups are short is to sample them heavily with a view toward breaking up any long branches. The best way to ensure that the branch connecting the two groups is short is to make the outgroup as close to the ingroup as you can, e.g. the sister group. But that’s a strategy aimed at breaking long branches, not one guaranteed to do so, and ultimately one has to work with the taxa and divergences that exist.
Now, if there’s any relevance to the various arguments about Xenoturbella, it’s this: Dunn criticizes others for preferring analyses in which choanoflagellates are the sole outgroup to Metazoa, claiming that additional, successive outgroups less closely related to Metazoa would be better. Absent the particulars of sampling and branch lengths, the expectation is that he’s wrong about that, for the reasons I have explained. Nobody, I think, is talking about adding outgroups that are closer to Metazoa or within Metazoa, so some of the questions you raise here don’t actually arise.
That’s the best I can do without knowing what you think is my problem.
Hi John, no I do not think you contradicted yourself – I admitted that I am very confused.
I thank you in advance for your patience.
You said: The best way to ensure that the branch connecting the two groups is short is to make the outgroup as close to the ingroup as you can, e.g. the sister group.
Yeah – I got that…
But do Chanoflagellates really fit the bill? Maybe they are closer to Metazoa than other (than what we once called single celled flagellated Protists) but maybe Choanoflagellates still remain too distant to be of much use on their own.
Bringing me back to an earlier line of inquiry above:
If you recall, you and I touched on this topic on Larry’s forum. Larry lamented (more than once) how different genes tell different stories and as far as he is concerned there is no unique ToL. Simon Gunkel & Georgi Marinov emphatically disagreed, at least for eukaryotes.
http://sandwalk.blogspot.ca/2015/05/ford-doolittle-talks-about-tree-of-life.html
http://sandwalk.blogspot.ca/2015/05/molecular-evolution-exam-april-2015.html
According to Georgi:
… such lineages begin to resolve better, and once you have many genomes, the picture improves considerably
maybe I am confounding phylogenetics and phylogenomics (I am out of my comfort zone) but I was thinking along the lines that if an adequate sister group is not at hand – alternate strategies that analyzed several out groups (not just the one) could still provide a better answer, but at the expense of exhorbitant computer time. I concede my naiveté.
At least that is how I understood Dunn… de-emphasis on “understand” as I repeat, this is well outside my comfort zone.
Thanks in advance for any help you could provide.
Maybe. But they’re the best we can get as long as relationships within Metazoa are in contention. Clearly, the sister group of whatever group includes Xenoturbella and all its possible closest relatives would be ideal, if indeed that’s the only question you’re interested in. The point is moot if we don’t know what that group is or what its sister group is.
No, that’s just wrong. The expectation is that several outgroups will be worse than a single, densely sampled outgroup (and by densely sampled I mean lots of species). Sheer number of species is not the most relevant variable. Again, the goal is to avoid long branches. Adding successively more distant outgroups just multiplies the number of long branches. Once more: it isn’t just “more genomes”; it’s more genomes that go in the proper places to reduce the lengths of tree branches. If a new genome actually increases the lengths of branches (because it’s outside all the other branches on the tree), that makes things worse, not better. And on this point there is no difference between phylogenetics and phylogenomics.
John Harshman,
Hi again John,
THANK YOU!
I remain in your debt.
I guess my only remaining question: How did Dunn’s paper pass review, on this score?