Hennig XXXI (23-27 Jun 2012) Riverside, California, USA

This year, the Hennig meeting was held at Riverside, California, and I'm fortunate enough to be able to be there (in great part thanks to the Kurt Pickett award, given by the Willi Hennig Society). As in previous occasions, I will try to give a short review of many of talks!

23 jun

I want to remark the excellent reception given by the organizers, and the opportunity to meet some wonderful heteropteran researchers [which is huge for an heteroptera aficionado! :)]

24 Jun

The meeting start up with a plenary talk given by Quentin Wheeler, with the title of Cladistic cosmology. Basically it is about of building a “Hubble like” project for taxonomy, specifically the description of new species, and facilities to make access to museum specimens across the web, both as metadata, but using some form of virtual specimens (e.g. scans of herbaria types, 3d scans of specimens, etc.). I like his point, but I think he miss the point of sampling: we not only need better understanding of current museum data, but also, we need to go to the field and look for never seen before specimens!

The first symposium was organized by Ward Wheeler about linguistics. The first talk, was a video-conf given by Peter Whitely about  the phonetic character definition used for an analysis of Uto-Aztec languages. The actual analysis was presented by Ward, that shows some of the particularities of using phonetic characters in comparison with DNA data, such as having 121 letters (against 4) and short strings (against long sequence fragments). Peter Forster talks about the using of phylogenetic networks for language data, and how it can be associated with what we know about the history inferred from mitochondrial molecular markers. You can found his program here. Joanna Nichols also uses phylogenetics networks with emphasis of using different sources of cognate data. To finish the season, John Wenzel shows a reanalysis of some previous indo-european analysis, showing some problems with the character coding, and how much of the current resolution was a by product of the errors in the codings, and the method used. I do not know much about the state of the field of phylogenetic linguistics (although I like the use of these techniques in linguistics!), I found a nice blog that usually talk about phylogenetics in linguistics here.

The next session was on phylogenomics, and was organized by Gonzalo Giribet, so are heavily biased towards invertebrates, but for a frustrated entomologist like me, it was wonderful :D! First Gonzalo talks about how new advances in sequencing technology has increased the amount of available data for phylogenetics, and how the problem was shifted from phylogenetic analysis to orthology recognition among this huge amount of new sequences. Then Torsten Struck shows how the selection of outgroups affects the position of one most mysterious group of annelids, the Myzostomida. Sebastian Kvist, explores the genomics of the blood feeding proteins used by Hirudinid leeches, and the realtionships of their host bacteria, a group related with nitrogen-fixing bacteria found in plants! Kevin Kocot was interested on orthology detection using phylogenetic trees, so he implement a program to detect the best possible set of ortholog sequences. Björn von Reumont talks about the pancrustacean phylogenomics, and how the acquisition of newer data from Remipedia is key in the problem and that results shows than remipedians are more closely related to hexapodans than to more traditional crustaceas, he also hints the existence of some apomorphies for such group based on nervous system. Vanessa González talks about the relationships among Bivalbia, that as many molluscan taxa, seems to be well known, but their phylogenetic relationships are just barely explored.

Then the poster season starts, most of the posters present a lot of entomological work on heteropterans (specially reduviids), so that makes me so happy, I like most the poster from Michael Forthman as he works with Ectrichodiins the group that I see for an small [very small!] amount of time xD, so hopefully he will construct a nice phylogeny of the group! :) The winning poster was given by Guayang Zhang and it is about the elongation of the front legs in harpactorines.

25 Jun

Students day! Most of students talks where given this day, so it will allow the judges to give the student awards with a complete overview of all the material presented by students.

Ansel Payne gives a wonderful talk about the general relationships of Hymenopteran superfamilies. Rebecca Dikow explores a huge data set (in term of characters) using bacterial genomes, so her trees have brutal lengths over 37 millions of steps! Ross Mounce talks about measuring support using supertrees, he also remind us about the importance of using machine readable data for phylogenetics, and that he thinks that using svg files will be great for publishing data, as you can include a lot of non-published metadata inside the files. Julieta Gallego presents a talk about the stability of molecular clocks (as I am a co-author of the talk, maybe I will talk about this in other time!). John Denton gives a talk about the usage of iterative alignments to improve the searches under likelihood approaches. Greame Oatley speaks about a cool group of birds [give me a passeriform, and I will love it ;)], and their species limits, in a work in part phylogeographic, in part taxonomic. Juanita Rodriguez (from Colombia! But she is working in Utah), talks about the phylogeny and biogeography of a group of Pompilid wasps. Fernando Gelin talks about Polybia, a wasp genus. Federico López (another colombian! And old friend of my from undergrad days! He is now at Vermont) gives a talk about the phylogenetics of a group of Vespids. To finish the first student season, I give my own talk about biogeography of amphibians.

In between student talks, there is an interesting symposium about the species problem. I'm personally no see so much in that problem, but is nice to see a lot of talk about that. Fortunatelly, I guess, although the problem is recognized, most taxonomists and systematist can continue to work, and work well, even if there is no consensus in what an species is. Brent Mischler spokes about the phylogenetic (topologic-monophyletic) species concept. Rudolf Meier talks whats about the problem of discussing the topic, and how most people skip the problem at all, and why he thinks it is important to discuss the subject. For me, this is the best talk of the symposium. Finally Kipling Will talks about the influence of the species definition in alpha-taxonomy, specifically how it is affected by some recent approaches, like bar-coding. Then a nice discussion (which includes Quentin Wheeler) was given by the four panelists of the symposium.

The second student seasson begin with a talk given by Christine Hayes about the the phylogenetic relationships of a group of phorid flies. Anna Dal Molin explore several techniques to visualize large groups of trees. Pei-Luen Li talks about a worldwide distributed (but concentrated in Hawai'i) group of asparagales, and Cecilia Waichert explore some adaptation hypothesis about nesting behavior using a phylogeny of Pomipilid wasps.

26 Jun

Pablo Goloboff, gives his talk about the implementation of Iter-PCR, focusing on the effects of changing the ingenuous algorithms, with the use of algorithms that take previous information into account, and makes feasible the use of iter-PCR in TNT. Jean DeLaet, gives an insight on his future program for phylogenetic analysis, Anagallis, that will include his algorithms for dealing with non-applicable data! I'm very curious about it [hopefully it will be open source!]. Nico Franz speaks about using taxonomy ontologies, although I think that kind of work is nice to discover some particular things, I guess that the missing thing is that they at the moment are no using character data, and I think that the key for this kind of databases will be both explicit references, and explicit data supporting each taxonomic affirmation. Lenka Drabkova talks about the plant cytokinins. Claudia Szumik shows the results of a the search of areas of endemism, using also phylogenetic information, with a huge data set of mammals. Then Santiago Catalano presents GB-to-TNT a user friendly program to build huge matrices, and also some nice taxonomy mapping tools found in the newer version of TNT. Daniel Janies, shows that super-trees are not a real solution to the uses of super-matrices, as they fails to produce approximate results of a super-matrix. In one of the most controversial talks Donald Buth argues that mitochondrial data, as non intrinsic source of data, and susceptible of the host-parasite interactions, must not be used for phylogenetic analysis. Personally I do not agree with him, but a friend of me, tells me that the same concerns have been raised in population genetics (a.k.a. Phylogeography) literature. So maybe, mitochondrial data is not good for small taxonomic levels, by the same reason of gene vs species trees, but at large levels, it will provide a nice source of evidence (as the difference between gene and species trees at this levels are no so important). This is an intriguing line of though, as it is precisely the opposite to the most received view, from at least five years ago. John Wiens gives a review of all of his papers about the use of phylogenies for study ancestral ecology. I was too tired after Jon talk, so I miss the next ones. Later Santiago gives a nice talk about  the usage of morphometric data on several and different kinds of studies. Tim Crowe talks about the results of a whole life dedicated to study guinea fowls at different taxonomic levels. Dalton Amorim, shows a new analysis of phorid flies based on morphology, one of the few talks using explicit morphological data! Continuing with morphological data, and dipterans, Torsten Dikow shows its phylogenetic analysis Mydidae, he also shows interest on detecting unstable terminals. Efraín de Luna, talks about the use of morphometric data in phylogenetics. To finish the day, Nico Franz gives a talk about how his analysis of some particular group of curculionids evolve in time from the initial data matrix to the definitive one. I think this is a nice exercise, but ultimately, it is difficult to extract some information about it, except from the anecdotal experience.

Later this day was the banquet, that as is a tradition in the hennig meetings was free for students! The banquet speech was given by Dennis Stevenson, the current editor of the journal of the society (Cladistics), and it was funny, although it was somewhat small!

The winners of the awards? Ok, there is a lot of Marie Stoppes and Kurt Pickett awards for traveling students. The Don Rosen award for the best poster was given to Guayang Zhang, as I tell you before. Cecilia Waichert receives the Lars Brundin award, for the second best talk, and John Denton receives the Willi Hennig award for the best student talk! Congratulation to the winners, all are well deserved, and I think the judge have a difficult time selecting these three, because there are a lot of great talks given by students :)!

27 Jun

I think this is the most difficult day to give a talk: everyone is tired of the amount of talks in previous days, as it is the day after the banquet, a lot of people are dehydrated and overtighted, and as it is the last day, you have to check-out from the hotel...

Here this is the last symposium, this one, about molecular clocks organized by Sean Brady, who discuss several sources of error. Tracy Heath talks about how to model dating in bayesian analysis, and Elizabeth Murray has a more concrete talk on the dating of some group of Hymenoptera.

Robert Sansom presents some differences between results using hard and soft anatomy in vertebrates, and the implications for the fossil record, that is only based on hard parts. Mark Simmons shows some problems from using multiple matrices, with small overlap. He present a lot of interesting results for some particular examples, so it is important to known how general they are. Mari Källersjö gives one of the most beautiful talks of the meeting, about a working project of a particular group of plants. Everyone loves his talk :D. Veronica Pereyra talks about thysanopterans, but unfortunately I have to make the check out, so I lost it :-(. Ronald Clouse talks about the phylogeography of an opilion that lives in southeastern US, that is a remaining of a gondwanian clade (as Florida was long time ago, part of African plate!). Ulf Jondelius, talks about the acoelan worms, a poorly known taxon, that he was sampled across the world. To finish the meeting Steve Farris talks about the recent surge of the three taxon statements authors (Ebach, Williams, etc.).

General overview

The meeting was very well organized by John Hartley and Christiane Weirauch, that move around all the time to make sure everything is right. They have a full group of students that are helping with everything, so it was huge :).

On the meeting itself, it has a lot of students, with several very nice talks, and a lot of discussion, that is always welcomed :D. One of the things that I note, is in the same line of previous meetings in which few morphology is used, here I see that although most studies only use molecules for phylogenetic analysis, people is more interested in link several of their findings with well detailed morphological data, hopefully, in the next few years, fully integrative, genomic and morphological data will be used in conjunction in almost every work!

The next year, the meeting will be held in Rostock, Germany, a city on the coast of the baltic sea. Hopefully I will be able to make it, and several of my cladistists ends will be there :D Also, if any one reading this blog interested in phylogenetics, and living somewhere in Europe, I encourage you to go, even if you don't work with parsimony! [In fact, a lot of people that does not use parsimony analysis shows their results at the meeting! And nobody complains about it, at much, someone as why do you not use parsimony, but just that] So! See you there :)!


Acknowledgments
I receive a lot of funding support from several institutions that make this trip possible: CONICET, FONCyT, GBIF (well, this one in the future) provides monetary funding, whereas INSUE allows me to work at their facilities. I receive a Kurt Pickett travel award from the Willi Hennig Society. At personal level, my advisors (Pablo and Claudia) provides me a lot of support, Julí convince me to go, Santí helps me with the trip logistics, and they, with Ross, they are excellent room-mates :)!

Hennig 30, July 29-August 2, 2011, Sao José do Rio Preto, Sao Paulo, Brazil

Just a week ago I was at the 30th meeting of the Willi Hennig Society (my third in four years! :D). It was in Sao José do Rio Preto, an small city in the state of Sao Paulo, Brazil. There is a lot of people! (More than 200! most of them are students!). Here I will try to give an small review of the whole meeting :) [In the same way as I do in previous ones ;)]

July 29

This was the day of the welcome party, I meet with some good old friends and new people. Also, I'm very tired after a long trip from Campinas to Sao Jose xD.

July 30

The real meeting starts here, with an small presentation of the presidency of the Willi Hennig (by Rudolf Meier) and the organization committee (Dalton Amorim and Fernando Noll). Then a talk by Mario Viaro about the evolution of the prepositión “ate.” It is nice to see some strong similarities between cladistics and linguistics! But I think that most of the talk is based on more or less informed speculation, I hope that new quantitative approaches (like the one show by Ward Wheeler) will invigorate the field!

The first symposium was about bioinformatic tools for the analysis of diseases. Dan Janies, comment of a new way to dealt with they web services of SupraMap, and how it evolves from a monolithic application, to a more piecemeal approach (as unix/linux users know it), it will be more flexible and then more useful (And as a developer, a lot easier to maintain!). If you are curious, its new web site is GisBank.

Then Julián Villabona, a long time friend, give a review of the different tools to understand the geographic spreading of dengue, based on his own previous and current work. I think that the nice lesson is the importance of using new quantitative approaches from historical biogeography, in this clearly phylogeographic framework.

There are two more talks on this symposium, but they are more orientated in bioinformatic questions (data mining, and so), a subject that I do not really feel, so I can take anything about it :P.

The second symposium is on Phylogenetics and conservation. The first talk, of course, was by Dan Faith, that review his PD measure, that I think is the best way to take phylogeny into account in a diversity framework. He shows some way to expand the measure to complementarity, I personally do not like that approach, as it miss one of the most robust things of the PD measure: that is independent of the sampling.

I skip Roseli Pellens talk, but I see the talk of Ronald Clouse about Schizomida of Micronesia, both in terms of phylogeny of the group and at population levels. He found some particular things about its distribution, as it seems that the explanation of its distribution is based on long dispersal, but they are very well separated even inside the small islands. I will be looking forward for his publication on this particular data set :D!

There is a poster season here, but there were a lot of posters, and most of them on very technical issues of particular groups, which is wonderful :D (I love to see biodiversity research) but I'm very ignorant on most of that groups :P

July 31

The third symposium was on biogeography, specially, neotropics. Dalton Amorim shows how the “relationships” among areas shows that there are different historical components on the neotropics, each with different times. Although I don't like the methodology of his work (I'm anti “cladistic biogeography”), I guess that their own study shows that looking for hierarchical patterns in biogeography is not the right route!

Eduardo Almeida talks about the breakup of Gondwana, and the alternatives. I feel he gives much weight to the “expanding earth” hypothesis. It is important to note that this is a fringe hypothesis in geology, in fact, it is almost never mentioned in books on tectonics :P. Actual plate tectonics is one of the our most powerful theories, so the expanding earth require a more compressive mechanism set. Sadly all the discussion on the expanding earth eclipsed the most interesting part of his talk, that is about the biogeographic patterns of Colletidae bees.

To end the symposium, Juan Morrone gives a review of some of the american transition zones, in Mexico and in Patagonia. This prompt an interesting discussion, and I feel, like in the talk of Amorim, that their own research shows that the hierarchic classification in biogeography is not feasible.

Then contributing papers season start. I was very afraid (there is a lot of people xD), but fortunatelly all the discussion on Juan's talk gives me a nice time to keep my mind and provide a nice starting point for my talk ;).

Then Cyrille D'Haese talk about the relationships of a beauty group of colorful and “giant” collembolans of New Caledonia and South west Asia. Fernando Dagosta discuss the relationships of a characid group of fishes (I guess that you can ask Marcos about this talk :P jejeje). Jerome Murienne talk was about niche modelling, and how to integrate them into a phylogenetic framework. I think that this is a very difficult topic, and I was gland that Jerome was very cautions on its presentation ;) as more approaches to integrete phylogeny and distribution modelling are brutally ad-hoc. Denis Machado talk was about the phylogeny of freshwater stingray's tapeworms.

Wei Song Hwang gives a talk about the phylogeny of Reduviidae (specifically Reduviinae subfamily, a non-monophyletic assemblage). It was a terrific talk, on a group that requires an extensive revision although I will prefer more morphological data to be coupled with the molecular data. Anyway, I love this talk. [Side note, I guess it helps that bugs are theo only group that I was on the verge to work with it xD]

Marcos Mirande shows its update of its phylogeny of Characidae--there is a lot of characid papers in this meeting! :)--with new characters, new taxa, and molecular sequences. Owen Davies, who I meet in South Africa, study a group of small african birds of the genus Cisticola, and how old taxonomic revision was good and also wrong, I like this appreciations for old works :D. To close the season, Guanyang Zhang talk about the phylogeny of Harpactorini, a tribe of neotropical Harpactorinae, another grouop of assassin bugs, that, if you live in south America, you can see waiting on leaves with their front legs raised into the air: they have some stiky glands to capture prey, and that was Guanyong main subject.

Another poster season, I just have few time to check it out the other posters :-/, as I have my own, and I was very busy there: I have a pair of very good discussions ;).

August 1

There is another symposium, this one about techniques for molecular phylogenetics. Lone Aagesen explore some different weighting schemes to dealt with full genomic data and explore some groupings detected under different gene duplications among angiosperms. Torsten Dikow about the effects on combining morphology and molecules in Asilidae, as in many cases morphological data set produce an excellent retrieval of actual classification, whereas combined data does not.

Then Ross Mounce give a great talk about how ILD was abused in the literature, and why the way in it was presented in publications made most of the results difficult to repeat--If you read Ross's blog, you will know that he is, rightfully ;), obsessed with result reproduction!--He also shows how there are several ways to extract information on the results of this test, beyond the probability of the test.

As I mentioned before, Ward uses the dynamic homology framework to explore the phylogenetic relationships among uto-aztec languages, I think that this approach provides a real improvement on the analysis of linguistic data, and I am waiting for its publication :D!

Then Mario DePinna gives a talk about ontogeny and rooting. I guess that this talk is most on the same line of the works Nelson and Platnick, that I despise, and I thing are hot topic about 20 years ago, but not now.

Pedro Romano was interested on the use of ordered and unordered multistate characters, and how that decision is not usually well explored in most phylogenetic works. Then Apurva Narechania shows a methodology called “Radical” to explore the data concatenation. I think that his approach can be expanded in several ways, for example, looking for stability after adding taxa (instead of characters), that I feel, is a unexplored field.

Tim Crowe compare the evidence value of nuclear vs. mitochondrial molecular data, and show how their behavior is just not as usually supposed (mitochondria work well on tips, nuclear data work well on deep nodes) in his extensive data set of Galliformes.

Herbert Ferrarezzi explore a way to code polymorphic taxa, instead of using single specimen molecular sequences. Jaime Rodriguez present gives a talk about the use of spectral firms to identify insects, he tries to link it with phylogenies, but I really do not understand how this was done.

The talk of Claudia Szumik was about new character systems used in the classification of Embioptera. I like her work as it is a clear example that in many groups, the “lack” of informative morphological characters, rest on traditional grounds (i.e. taxonomic research centered on some particular character systems), rather than in a really uninformative morphology.

To end the day, Gustavo Hermes speak about a difficult wasp group, the Eumeninae.

The banquet!

This day was the banquet, in a place with a lot of food of several sources: fish, shrimps, sushi, meat (a lot!). It was great ;). And of course there are the Banquet speech (about “addiction on cladistics,” an excellent talk given by John Wenzel), and the Society awards for students, both the travel grants (Marie Stoppes) and the Don Rosen award for the best poster (to Rafaela Lopes), the Lars Brundin award for the second student talk (to Gustavo Hermes), and the Willi Hennig award for the best student talk (to me! I was pretty surprised :D, and really happy! :D). Congrats to the winners :), I'm proud of be in such nice company :D!

August 2

I must admit that I was pretty tired after two days of brazilian joy ;) so, I just barely play attention to several of this talks :P, and loss the talks from Phillippe Grandcolas and Nobuhiro Minaka.

Hilton Japyssu talk about how to codify different behavior patterns, in particular grooming on rodents. They produce a very well supported phylogeny with a lot of this rutines, congruent with several other data sources. Then Carlos Alberts, on the same line work with behavior patterns of Arini parrots.

Later, Santiago Catalano present his method about landmark alignment, and how it produces better results than alignments that not take into account the phylogeny. Another hit for the dynamic homology approach.

One of the talks I want to hear, is the one by Jan DeLaet on character weighting. He was very interested on implied weighting approaches, but it seems that he prefers linear functions instead of concave ones, so he present some properties for a particular function schema that he call, “self calibrated.”

To finish the meeting there is a symposium on theoretical and phylosophical aspects on cladistics. The first talk was by Pablo Goloboff about some improvements to implied weights, that is, to use different weighting functions among different characters, and the possibility to weight blocks of characters instead individual characters (that is mostly for molecular data).

James Carpenter gives a talk about the meaning of homology vs. synapomorphy. Of course, synapomorphy is not homology, as you can also have homology in symplesiomorphy! In the same line Kevin Nixon (who was not there, but send a video of his talk) was on the same line. This might be an old fashion discussion, but given that there is a lot of recent publications by people like Malte Ebach, David Williams and company (they publish a book, and have several papers in journals like zootaxa) it is important to recall this arguments again.

Andy Brower, tries to show how really most of objections on parsimony are not so strong, and that actually, parsimony is not as bad as their criticizers argue: the results are nearly identical with alternative methods. There are several topics I agree with Andy, but there are others that not, mostly he uses some sociological explanations in some cases, but ignore them in others (and of course, I'm very realist :P ejeje, but that is my own problem xD).

The second talk of Jan was about the topics touched on its 2005 paper, that is, about the use of gaps weighting and how they are related to dynamic homology, and of course, why really using everything in 1, is not really a straight forward consequence of using parsimony.

To finish, Steve Farris criticize the most recent attempts of users of 2 taxon analysis approach (already mentioned: Ebach, Williams, Nelson, Platnick), and how they change their own words when new criticisms where made.

Overview

A very long piece here :P! It was a very nice meeting. With a lot of students, that as great, and I think the most important objective of the WHS: to bring students into phylogenetics. But there is disappointment, apart from brazilian students there are few from other countries. I guess that it is a consequence of recent economic downside, as there are just few US students (3 or 4), and just one European students (just Ross). It is shocking when you see such small representation from first world countries. Representation from latinamerica is also scarce: just one from Argentina (were was everybody?), and nobody from other latin American countries. Of course, there are some colombians working on Brazil, but they count as Brazilian students (as its fundings are given by brazilian institutions). I don't know what happens, it is not lack of publicity: there is the same amount as in any other Hennig meeting. It is not the price, after all, this is one of the cheapest international meetings out there!

Nevertheless, that is not a problem of the organization, so I just can tell that the meeting was great, a lot of students, a nice people, and several interesting talks and posters :) Now, it is time to think on the next one, which will be somewhere in the U.S., hopefully I can made it :D!

Acknowledgments

Of course, a lot of fundings is required to do this trip, and I want to acknowledge the support given by the FONCyT, CONICET and the INSUE, and the Willi Hennig Society that gives me a Marie Stoppes award, and the Organization committee that give us a discount rate at the hotel, and everyday lunch ;)! Obrigado! :D

Hennig XVII: “Live” blogging, day 2

Today the meeting starts with a highly molecular morning. Gonzalo Giribet present a symposium about new methods for “phylogenomics” (organized by him, Ward Wheeler, and Jirky Mouna). Three talk were about the use of gene order, inversions, recombination, in the context of phylogenetic analysis. All of the presented analysis are POY-oriented ;). It is sad that Andres Varón, a colombian working with POY, was unable to assist.

The most interesting one, is the talk of Gonzalo who insists the usefulness of morphological data, and present some new analysis with his EST data for metazoa (published in Dunn et al. 2008), with more than 200 morphological characters for metazoans in a network of experts. His analysis show that morphology provide evidence for grouping at all levels of the tree. He was very suspicious about the “groundpland” coding. I always prefer exemplar coding, but sometimes, some useful information available from not directly analyzed terminals (for example, several developemental data) can provide an excellent source of information.

Prosanta Chakrabarty, try to test sexual selection in a group of luminous fishes, I get impressed with the diversity of that fishes, but, I think that the questions about selection preasures can not be answered in the way showed by Prosanta (or in any way!).

Next, there are two nice works on Curculionids. In the first Analía Lanteri, showed a particular group of broad nose weevils, then Adriana Marvaldi, showed his most recent advances in the understanding of the whole morphology (and phylogeny) of Curculionoidea, and how many of the sequence data recently assembled for that beetles, is highly congruent with the morphological results.

Afternoon talks are more interesting to me, because they are about biogeography ;) --I only have 3 interests: parsimony methodology, morphological phylogenetics, and REAL biogeography xD--.

In the first one, Peter Hovenkamp, shows a very interesting parallel between ideas form phytosociology and endemism, he founds that many of the implementations of Josias Braun-Blanquet (or the europoean school of phytosociology really did not implement that methods! He thinks that NDM (of Szumik and Goloboff, 2004) can be a better tool for phytosociological ecology!

Another wonderful talk was giving by Claudia Szumik, who made an analysis of endemism of northern argentina. The good thing about the study, is that it includes data taken directly from the experts of each group, with several data collected by Claudia and their co-authors!

Loló (Dolores Casagranda) present a comparison between NDM and PAE, I work on that talk, so I feel that a I will give a biased report, so I pass xD.

Erika Parada, an student from my undergrad university, talks about his analysis of northern andes, she uses tree-fitter,  a program that I do not like a bit :P, because it has several problems, I think she made a great work, but unfortunatelly results with TF are, for me, doubtful! :P

I do not like the talk from Dalton Amorim, but the discussion that follows the talk was very interesting, with James Liebherr, given several strong (and clever!) points against Dalton ideas.

“Weighting” tress with TreeFitter

TreeFitter [1] is a program for matching phylogenies with associations used in biogeography and co-evolutionary studies. It has some problems, as seems that Treefitter move to open-source, maybe this problems can be solved! Here I address some problems produced with the 'weighting tree' procedure. The analogous problems is found in some phylogenetic methods and programs, and I address it laterally (they consequences are fully discussed in [2]).

As many biogeographic programs [3, 4, 5] TreeFitter only dealt with perfectly dicotomic trees. To overcome this fault, it implements a weighting of trees. Then you can put all the dicotomic trees found in the phylogenetic analysis and put a fractional weight to each tree. For example, if you found four trees, each one would be weighted by 1/4. This is in fact a majority rule consensus. Ronquist, who is a defender of bayesian methods, see the weighting of trees as a positive characteristic as it covers the 'uncertainty' of the analysis [6]. Then it express the 'confidence' (support in a most relaxed version) of each clade. But contrary to intuitive expectations, majority rule consensus have nothing to do with the support of a determinate clade. Instead, they favoring ambiguous topologies! [2, 7].
Take this example (after [2] and [7]):

#nexus

ptree new1 weight=0.143 (1,((6,(7,(8,9))),(10,(4,(3,(2,5))))));
range new1 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new2 weight=0.143 (1,((6,(7,(8,9))),(2,(3,(4,(5,10))))));
range new2 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new3 weight=0.143 (1,((6,(7,(8,9))),(2,(3,(5,(4,10))))));
range new3 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new4 weight=0.143 (1,((6,(7,(8,9))),(2,(3,(10,(4,5))))));
range new4 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new5 weight=0.143 (1,((6,(7,(8,9))),(2,((3,10),(4,5)))));
range new5 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new6 weight=0.143 (1,((6,(7,(8,9))),(2,(10,(3,(4,5))))));
range new6 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree new7 weight=0.143 (1,((6,(7,(8,9))),((2,10),(3,(4,5)))));
range new7 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i, 10:g;

ptree indie1 (1,((6,(7,(8,9))),(4,(3,(2,5)))));
range indie1 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i;

ptree indie2 (1,((6,(7,(8,9))),(2,(3,(4,5)))));
range indie2 1:a, 2:b, 3:c, 4:d, 5:e, 6:f, 7:g, 8:h, 9:i;

htree host1 (a,((f,(g,(h,i))),(d,(c,(b,e)))));
htree host2 (a,((f,(g,(h,i))),(b,(c,(d,e)))));

The ambiguity is caused by the taxon '10' of data set 'new' that jumps to several positions among the tree. '10' has not influence on the selected tree because it is a product of a dispersal in all topologies. '10' is inestable in all topologies that include (4,5), then the majority rule gives more weight to that topology than to alternative topology, in which '10' position is not ambiguous. In this case both 'tree islands' have the same evidential weight (by the way, that is the reason to prefer strict consensus over other consensus!). But when weights are applied the topology showing (4,5) are preferred as there is more topologies with that clade, then the first tree is preferred because they lack of resolution!

At first look it seems that this case can be solved weighting the whole islands instead of each tree, but within each island it is possible to have the same problems of the first example, and in more complex cases, identifying 'topology' islands became difficult and maybe impossible if there are several combinations in independent clades!

Solutions?

Of course the best solution for the problem of multiple trees in TreeFitter without using weights is a new version that dealt with polytomic trees. I guess that the resistance against polytomic trees is because they 'imply' simultaneous speciation. I do not hold that kind of idea ;), and I have no reason to think that polytomic trees supports such interpretation. Even if that is the interpretation is better than weighting (by the way, if you think that polytomic trees implies multiple speciation, the weighting implies fractional speciation! I think that it is a more problematic idea than 'multiple' speciation!).

But before a new version of TreeFitter--or a similar program--arrives it is necessary to found a solution to the problem of multiple trees. I am not happy with the solution that I propose here, but I have no other idea, so here I go...

Use an Adams consensus to detect the terminal/clades that produce the multiple trees, and remove it form the analysis, so you keep the stable part of the topology. I do not like removal of evidence, but it seems safer than relying in the biased solution of tree weights. Maybe some want to re-run the analysis, but I think that is preferable to use the reduced tree as it is based directly on the whole evidence, then, the effect of data removal, I hope, is minor, as the stable part of the original topology is conserved. Also, TNT [11, 12] has several tools to identify moving taxons, so an analysis of the trees with TNT, provides several ways to found stable topologies.

A second problem is that in TreeFitter, the extinctions had a cost greater than 0 [6], so removing terminals could increase the extinction value, but I think that this is a minor problem compared with given weight to ambiguous data. Actually we can think that the cost increase by the new 'extinctions' is the penalty for the ambiguity of the data.

Weighting trees in other methods

As far as I know there are other phylogenetic methods/programs that weight trees, and suffer from the same flaws as the tree weighting under TreeFitter.

First, is the majority rule consensus. It seems that the main reason to prefer majority rule consensus is because they provide results with more resolution. But as examples shows [2, 7] the resolution created by a majority rule tree is coupled ore with the ambiguity generated by a particular topology skeleton. Maybe a more interesting solution could be the use of a minority rule consensus (note that supported clades, always appear as they minority rule is 100%), an advantage is that at least they knowledge that the number of clade instances are not related with the 'support' of the clade. A comparison between the majority and minority rule consensus can show the parts of the tree that are somewhat unstable. But I think that this kind of analysis is better performed using a combination of an strict consensus and an Adams consensus [8], or several of the tools from TNT [11, 12].

Under Bayesian analysis, the frequency of clades is recognized as a 'posterior probability' (a more catholic interpretation might be clade support). Bayesian analysis differs from typical consensus because they made a consensus from trees with different optimality value. But all explored topologies are taken as they are found, and branches are never collapsed, so they are subject of the problems of majority rule consensus [2, 7]. Then in cases where a particular terminal/clade is producing ambiguity, the final topology favors the ambiguous topology. This can produce some illogical results when the data sets even with a single optimal tree, are not very decisive [9], with several near-optimal fit trees. In that case, is is possible that the method prefer the ambiguous topologies from the sub-optimal trees, over the topology in the optimal one, even with high probability values! I recommend you to read [2] for a full criticism of bayesian analysis.

There is also other form of majority rule consensus is used: for support measuring using resampling (like jacknife or bootstrap). In this case the use of majority rule reflects the amount of support for each clade. First, from each resampled matrix analyzed a strict consensus tree is build, then the final majority rule consensus shows the amount of times in which a supported clade appears. In this case there is no bias against or for a particular clade, because the trees from resampled matrix are collapsed. This is not the case for PAUP [2, 10], because in PAUP the trees are weighted in each resampled search, then, the majority rule consensus is a majority rule consensus of several individual, and un-collapsed trees, then it is fully prone to the ambiguity problems of majority rule consensus.

[1] Ronquist, F. 2001. TreeFitter, program and documentation. Available at: http://www.ebc.uu.se/systzoo/research/treefitter/treefitter.html
[2] Goloboff, P.A., Pol, D. 2005. Parsimony and Bayesian phylogenetics. In: Albert, V.A. Ed. Parsimony, phylogeny, and genomics. Oxford univ., Oxford, pp. 148-159.
[3] Page, R. D. M. 1993. Component 2.0, program and documentation. Available at: http://taxonomy.zoology.gla.ac.uk/rod/cpw.html
[4] Page, R. D. M. 1994. TreeMap 1.0, program and documentation. Available at: http://taxonomy.zoology.gla.ac.uk/rod/treemap.html
[5] Ronquist, F. 1996. DiVa, program and documentation. Available at: http://www.ebc.uu.se/systzoo/research/diva/diva.html
[6] Ronquist, F. 2003. Parsimony analysis of coevolving species associations. In: Page, R.D.M. Ed. Tangled trees. Chicago Univ., Chicago, pp. 22-64.
[7] Sharkey, M.J., Leathers, J.W. 2001. Majority does not rule: the trouble with majority-rule consensus trees. Cladistics 17: 282-284. doi: 10.1006/clad.2001.0174
[8] Kearney, M. 2002. Fragmentary taxa, missing data, and ambiguity: mistaken assumptions and conclusions. Systematic biology 51: 369-381. doi: 10.1080/10635150252899824
[9] Goloboff, P.A. 1991. Homoplasy and the choice among cladograms. Cladistics 7: 215-232. doi: 10.1111/j.1096-0031.1991.tb00035.x
[10] Swoford, D. 1998. PAUP, program and documentation, Sinauer, Sunderland (USA).
[11] Goloboff, P.A., Farris, J.S., Nixon, K.C. 2008. TNT, program and documentation. Available at: http://www.zmuc.dk/public/phylogeny/TNT/
[12] Goloboff, P.A., Farris, J.S., Nixon, K.C. 2008. TNT, a free program for phylogenetic analysis. Cladistics, in press. doi: 10.1111/j.1096-0031.2008.00217.x

Dispersion strikes back... but in the wrong direction

Just two weeks before, a seminar about biogeography and phylogenetics starts here, in Tucumán. We were about 8 people taking about a biogeographic paper (the next week, it will be a phylogenetic one). So there are my own impressions about the papar (I posted it late, because I'm out of the town because of the of the holly week break).

Sanmartín, I., van der Mark, P., Ronquist, F. 2008. Inferring dispersal: a Bayesian approach to phylogenetic-based island biogeography, with special reference to the Canary islands. J. biogeografy 35: 428-449. DOI: 10.1111/j.1365-2699.2008.01885.x

Personally I do not like Bayesian methods in phylogenetics, they are based on several flawed principles [1], but I will try to focus in the other aspects of the proposed framework of Sanmartín et al., instead of attacking the Bayesian principles.

As many others, Sanmartín et al., attack 'vicariance' biogeography on the grounds that they only counts vicariance and ignore dispersal, I think that the tale is somewhat different, vicariance biogeographers are fully aware of dispersal, but they can't find grounds to found a general dispersal pattern, in the other hand vicariance provides a fully explicable pattern that is shared by several and unrelated organisms, then dispersal, can only be addressed group-wise. But in the case of the islands without any connection to land, vicariance can not be used as a general explanation. But can dispersal provide one?

Sanmartín et al., argue that the answer is 'yes', but instead of show why such conclusion arises they jump to their own model of dispersal, the causal mechanism that put several different organisms in the same model is never given, then their own asseveration that dispersal can produce 'concerted' (i.e. general) patterns, which is the most interesting question raised by they are never answered. What is striking is that they show several different models of dispersion proposed for canarian taxa.

Moreover the model developed, is a symmetrical one, then if in fact it is a common dispersal route, by an oceanic stream for example, it can be masked by their model that presupposes that both dispersal directions are equally probable.

Then they jump on the model frenzy, they are as embedded on the power of model approaches, that they raise 'new questions', that were irrelevant before their model. The most striking one is the 'carrying capacity' of their model. Carrying capacity is a term borrowed from ecology, and in this case is somewhat equated with island raw richness. This parameter, that seems to be originated from the 'island biogeography' of Wilson, can be of interest to ecologist, but they relevance in the search of common biogeographical patterns is never clearly showed (and I do not think that it has some empirical meaning), but now, this is an important question: most of Sanmartín et al. discussion, are about carrying capacities, not about dispersal models, which taxa are best fitted to the model, dispersal rates for different groups, or common routes of dispersal. Sanmartín et al. transform the parameter calibration into the research answer, losing the important questions in the middle.

When given the results, Sanmartín et al. contradicts many of its initial reasons to choose a likelihood model, the first one is that the likelihood of each model allows a selection of the models, but as the model (and its parameters) with the best likelihood seems to be illogical, they move on a sub-optimal models, they claim that it is not enough data, but when is enough data? How can I selecet a model only guided by its unexpected results? Why no limit the model to certain values if we are afraid of illogical answers? I think that Sanmartín et al. Results actually undermined their claims.

Sadly the paper is more like a discourse from a politician in campaign: all about the promises of wonderful results that might be open for likelihood models, but the actual use of the model with the empirical case is disappointing.

If a dispersal biogeography based on 'concerted' patterns of dispersal is to be used, their user would answer several questions: (1) why expect a 'common ' dispersal pattern? (2) is their model adequate to the desired answer? (3) Is the pattern only spatial (i.e. a common dipsersal route) or spatial and temporal (i.e. a common dispersal route, roughly at the same time)? (4) What about the taxa that departs from the selected model? Questions (1) is never answered by Sanmartín et al., question (2) is a clear no (the symmetrical model contradicts directly their aim), maybe they can give an answer to question (3) but they are more interested in parameter estimation than in biological questions, the question (4) is never answered, even when they accept that there are several different models proposed for different taxa in the Canarias.

[1] Goloboff, P.A., Pol, D. 2005. Parsimony and Bayesian phylogenetics. In: Albert, V.A. Ed. Parsimony, phylogeny, and genomics. Oxford univ., Oxford, pp. 148-159.