Pierre Pontarotti Editor

Evolutionary BiologyConvergent Evolution, Evolution of Complex Traits, Concepts and Methods

Evolutionary Biology

Pierre PontarottiEditor

Evolutionary BiologyConvergent Evolution, Evolutionof Complex Traits, Concepts and Methods

123

EditorPierre PontarottiCNRS, Laboratoire Evolution Biologique etModélisation

Université d’Aix-MarseilleMarseilleFrance

ISBN 978-3-319-41323-5 ISBN 978-3-319-41324-2 (eBook)DOI 10.1007/978-3-319-41324-2

Library of Congress Control Number: 2016942867

© Springer International Publishing Switzerland 2016This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or partof the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmissionor information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilarmethodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exempt fromthe relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, express or implied, with respect to the material contained herein orfor any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer NatureThe registered company is Springer International Publishing AG Switzerland

Preface

For the ninth year we publish a book on evolutionary biology concept and appli-cation we try to catch the evolution and progress of this field for this goal we arereally helped by the Evolutionary Biology Meeting at Marseilles. The goal of thisannual meeting is to allow scientists of different disciplines, who share a deepinterest in evolutionary biology concepts, knowledge and applications, to meet andexchange and enhance interdisciplinary collaborations. The Evolutionary BiologyMeeting at Marseilles is now recognised internationally as an important exchangeplatform and a booster for the use of evolutionary-based approaches in biology andalso in other scientific areas.

The book chapter have been selected from the meeting presentations and from aproposition born by the interaction of meeting participants.

The reader of the evolutionary biology books as well as the meeting participantswould, maybe like me, witness years after years during the different meetings andbook editions a shift on the evolutionary biology concept. The fact that the chaptersof the book are selected from a meeting enables the quick diffusion of the novelties.

I would like to underline that the nine books are complementary one to anotherand should be considered as tomes.

The articles are organised in the following categoriesConvergent evolution (Chaps. 1–8)Evolution of complex traits (Chaps. 9–14)Concepts (Chaps. 15–19)Methods (Chaps. 20–22)

Marseille, France Pierre PontarottiApril 2016

v

Acknowledgement

I would like to thank all the authors, the meeting participants, the sponsors of themeeting: Aix Marseille Université, CNRS, ITMO, ECCOREV FEDERATION,Conseil Départemental 13, ITMO, Ville de Marseilles.

I wish to thank the team of the non profit organisation: Association pour l’Etudede l’Evolution Biologique (AEEB) for the organisation of the meeting.

I also wish to thank the Springer’s edition staff and in particular AndreaSchlitzberger for her competence and help.

I am also thankful to the director of the AEEB Marie-Hélène Rome for thecoordination of the meeting and the book.

Marseilles France Pierre PontarottiApril 2016

vii

Contents

Part I Convergent Evolution

1 Road Map to Study Convergent Evolution: A Propositionfor Evolutionary Systems Biology Approaches. . . . . . . . . . . . . . . . 3Pierre Pontarotti and Isabelle Hue

2 Analysing Convergent Evolution: A Practical Guideto Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Kevin Arbuckle and Michael P. Speed

3 Convergent Evolution Within CEA Gene Families in Mammals:Hints for Species-Specific Selection Pressures . . . . . . . . . . . . . . . . 37Robert Kammerer, Florian Herse and Wolfgang Zimmermann

4 Convergent Evolution of Starch Metabolism in Cyanobacteriaand Archaeplastida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Christophe Colleoni and Ugo Cenci

5 The Evolution of Brains and Cognitive Abilities . . . . . . . . . . . . . . 73Christopher Mitchell

6 Convergence as an Evolutionary Trade-off in the Evolutionof Acoustic Signals: Echolocation in Horseshoe Batsas a Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89David S. Jacobs, Gregory L. Mutumi, Tinyiko Malulekeand Paul W. Webala

7 Convergence and Parallelism in Astyanax Cave-Dwelling Fish . . . . 105Joshua B. Gross

8 Evolutionary Pathways Maintaining Extreme Female-BiasedSexual Size Dimorphism: Convergent Spider Cases DefyCommon Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Matjaž Kuntner and Ren-Chung Cheng

ix

Part II Evolution of Complex Traits

9 Evolution of the BCL-2-Regulated Apoptotic Pathway . . . . . . . . . . 137Abdel Aouacheria, Emilie Le Goff, Nelly Godefroyand Stephen Baghdiguian

10 The Axial Level of the Heart in Snakes. . . . . . . . . . . . . . . . . . . . . 157J.W. Faber, M.K. Richardson, E.M. Dondorp and R.E. Poelmann

11 On the Neo-Sex Chromosomes of Lepidoptera. . . . . . . . . . . . . . . . 171Petr Nguyen and Leonela Carabajal Paladino

12 Recent Developments on Bacterial Evolutioninto Eukaryotic Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187Mauro Degli Esposti, Otto Geiger and Esperanza Martinez-Romero

13 Genomic Analysis of Bacterial Outbreaks . . . . . . . . . . . . . . . . . . . 203Leonor Sánchez-Busó, Iñaki Comas, Beatriz Beamud,Neris García-González, Marta Pla-Díazand Fernando González-Candelas

14 Three-dimensional Genomic Organization of Genes’ Functionin Eukaryotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233Alon Diament and Tamir Tuller

Part III Concepts

15 How Likely Are We? Evolution of Organismal Complexity . . . . . . 255William Bains

16 Molecular Challenges to Adaptationism . . . . . . . . . . . . . . . . . . . . 273Predrag Šustar and Zdenka Brzović

17 Ontogeny, Oncogeny and Phylogeny: Deep Associations . . . . . . . . 289Ramray Bhat and Dharma Pally

18 Separating Spandrels from Phenotypic Targets of Selectionin Adaptive Molecular Evolution . . . . . . . . . . . . . . . . . . . . . . . . . 309Stevan A. Springer, Michael Manhart and Alexandre V. Morozov

19 From Compositional Chemical Ecologies to Self-replicatingRibosomes and on to Functional Trait Ecological Networks . . . . . . 327Robert Root-Bernstein and Meredith Root-Bernstein

Part IV Methods

20 Inference Methods for Multiple Merger Coalescents . . . . . . . . . . . 347Bjarki Eldon

x Contents

21 From Sequence Data Including Orthologs, Paralogs,and Xenologs to Gene and Species Trees. . . . . . . . . . . . . . . . . . . . 373Marc Hellmuth and Nicolas Wieseke

22 Oh Brother, Where Art Thou? Finding Orthologs in the Twilightand Midnight Zones of Sequence Similarity . . . . . . . . . . . . . . . . . 393Bianca Hermine Habermann

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421

Contents xi

Part IConvergent Evolution

Chapter 1Road Map to Study Convergent Evolution:A Proposition for Evolutionary SystemsBiology Approaches

Pierre Pontarotti and Isabelle Hue

Abstract Every evolutionary biologist will surely acknowledge that convergentevolution, the independent evolution of similar features in different evolutionarylineages, is an important phenomenon of the organic evolution. However, theconcept is complex and can have several related but conceptually distinct meaningsin the literature, including parallel evolution (independent mutations in orthologousgenes) and homoplasy (any convergent traits, including reversion). Some authors,for example, use the term “parallel evolution” differently from “convergent evo-lution”, and they reserve the latter term for more “unlikely” (or “more indepen-dent”) examples of phenotypic similarity across lineages, those not predisposed bygenomic similarity. Semantic arguments in science are often fruitful, but can alsoprevent efficient scientific exchanges in the field. Hence, the goal of this article was(1) to define convergent evolution in a better way by applying a multilevelbiological-level approach and (2) to propose a road map to help researchers navi-gate their routes in studying this phenomenon.

State of the art including new concepts needed to better understand thisphenomenonIn order to add depth and (we hope) some clarity to the conceptual understanding ofevolutionary convergence, it is important to describe the following aspects: at thephenotype level, to distinguish isoconvergent/alloconvergent evolution (see textbelow), species global convergence versus character convergence (multivariateconvergence vs single trait convergence) and identify what is the type of characterthat has evolved, whether it is morphology versus physiology, for instance, andwhether it is a continuous or discrete character.

P. Pontarotti (&)CNRS Centrale Marseille, I2M UMR 7373 équipe EBM (Evolution BiologiqueModélisation), Aix Marseille Université, 13453 Marseille cedex, Francee-mail: pierre.pontarotti@univ-amu.fr

I. HueUMR BDR INRA ENVA, Université Paris Saclay, 78350 Jouy en Josas, France

© Springer International Publishing Switzerland 2016P. Pontarotti (ed.), Evolutionary Biology, DOI 10.1007/978-3-319-41324-2_1

3

We argue here that it is essential to first characterize phenotype-level conver-gence, before defining convergence at other levels. Indeed, once the phenomenon atthe phenotypic level is clearly described, one can define the phenomenon at otherbiological levels in cases where the different biological levels are linked, andaddress the question about the origins of the genetic changes de novo versus thepre-existing ones and the consequences of the genetic variation at the coding,epigenetic, transcriptional and higher biological level.

1.1 Concept, Definition: Convergence at Phenotypic Level

1A) Iso- versus Alloconvergent evolution

We start this section with a short epistemologic analysis about the use of theterm parallel and convergent evolution in the literature. The term “convergentevolution” is generally used in the following way: «Convergent evolution: theindependent evolution of similar features in different evolutionary lineages» (Losos2011). However, the convergent evolution of a given character can result from theevolution of a similar or a different character.

The distinction, using the ancestral state information, is found in some of themolecular phylogeny analyses, under the term of parallel and convergent evolution(Zhang and Kumar 1997). Here, the term “parallel evolution” is used if the ancestralamino acid is the same, and “convergent evolution” is used if the ancestral aminoacid is different.

However, most of the time, when convergent evolution is found at the phenotypelevel, the distinction between parallel and convergent evolution is not based on theevolutionary history of the characters, but on the similarity of the genetic mecha-nisms that are involved in the repeated phenotype. If the molecular mechanisms arethe same, the evolution is said to be parallel; if the genetic mechanisms are different,the evolution is said to be convergent (see, for example, Rosenblum et al. 2014).Some authors use also the phylogenetic proximity; when the species are phyloge-netically close, the term “parallel evolution” is used; when they are phylogeneti-cally distant, they use the term “convergent evolution”. This makes sense, regardingthe above use of both terms, since closest species will tend to evolve using the samemechanisms, either because the variation that underlines the phenotype variationcorresponds to the standing variation (Elmer and Meyer 2011), or because of theinteraction between the different genes of the genome (or epistasis). With a similarbackground, more similar fixation of substitutions are excepted but if the interac-tions are different between the genes, it is likely that the evolutionary trajectory willbe different (for a great example, see Ujvari et al. 2015).

In order to help in the conceptual understanding of evolutionary convergence,we propose two neologisms that can be applied to all the biological levels: iso-convergent and alloconvergent evolution (iso from the same ancestral state and allofrom a different ancestral state). This distinction is not done at the phenotype level

4 P. Pontarotti and I. Hue

in the literature (Conway Morris 2003; McGee 2011; Losos 2011; Gordon et al.2015). However, it has to be noted that scientists who work at the genetic level,when talking about convergent evolution at the phenotype level, intuitively meanisoconvergence at the phenotype level (see Fig. 1.1 in Martin and Orgozozo 2013;Rosenblum 2014). As an illustration, this is what Rosenblum et al. wrote (2014) ina chapter called Linking phenotype and phylogeny: “Before investigating themolecular mechanisms of convergence, researchers must first ensure that thephenotype of interest is convergent. Researchers should define convergent evolu-tion in a phylogenetic context. This requires an explicit integration of phenotypicdata with a molecular phylogeny, and should incorporate uncertainty in the phy-logeny and the model. For example, the independent evolution of similar pheno-types can be identified from ancestral state reconstruction, comparisons ofphylogenetic and genetic distance, or inferred shared selective regime (e.g.Muschick et al. 2012; Ingram and Mahler 2013)”. Here, Rosenblum et al. used theword “convergent evolution” of a phenotype, but they mean, in our terms, iso-convergent evolution of a phenotype.

Thus, either the distinction is not noted, or the authors (mainly the authorsinterested by the genetic mechanisms) when discussing about convergent evolutionthink about “isoconvergent evolution” implicitly. In our view, it is, however,important to explicitly make the distinction between iso- and alloconvergence at thephenotype level as, above all, this will allow testing further the relationship betweenisoconvergent evolution at the phenotypic level and isoconvergence at other bio-logical levels (Stern and Orgozozo 2008). If this is the case, isoconvergent evo-lution of a phenotypic level can be used to decipher biological mechanisms at thegenetic, epigenetic, transcriptional and other biological levels (Kopp 2009).

A key example: biochemical/enzymatical functionAlloconvergent evolution of enzyme function can be seen in two distinct, but some-times joint, effects. The first is when non-homologous enzymes deliver the sametransformation, as expressedby the same four-digit enzymecommission (EC)number.These enzymes are named transformational analogues. The second situation is whenthe same (same four-digit EC) or related (same three-digit EC) enzyme transformationis effected by a similar disposition of residues in the active site, as exemplified in theSer-His-Asp catalytic triad shared by the trypsin family and subtilisin. Such enzymesare mechanistic analogues. These two situations are not exclusive because twoenzymes are assigned to both classes if they perform exactly the same overall reactionwith the same mechanism (Doolittle 1994; Gherardini et al. 2007).

The transformational analogues evolved at the protein level from differentancestors via divergent evolution as the transformational function evolved likelyfrom different ancestral transformational functions, and these functions evolvedtherefore in an alloconvergent manner (see, for an example, the case of b lactamaseB, Alderson et al. 2014).

In the case of enzymatic isoconvergent evolution, the same function is present inthe ancestor (see, for example, Dick et al. 2012).

1 Road Map to Study Convergent Evolution: A Proposition … 5

A

(a)

(b)

B

B

B

B

B

A

B

B

B

B

B

B

B

B

B

B

B

B

A A

B

A

Fig. 1.1 Isoconvergence versus alloconvergence. Legend: A: Derived state; B and C: Ancestralstate

6 P. Pontarotti and I. Hue

1B) Assumption of “convergent evolution” via adaptation to a newenvironment

Another aspect of the definition of convergent evolution is that some authorshave no assumption concerning the evolutionary driving force, while others see theaction of the adaptive evolution. Convergent evolution, defined as being the resultof adaptive evolution, is commonly used in the literature as synthesized by Stayton(2015). A scientific approach should be first to evidence convergent evolution andthen test for adaptation or any other phenomenon such as constraint or evolutionaryrelaxation (Arbuckle et al. 2014). Indeed, in many articles, the convergence at thephenotype level is suspected by the fact that the species that live in the sameenvironment will be adapted to it in the same manner and will have the samephenotypic response, and the case of the marine mammals is one of the examples(Foote et al. 2015; McGowen et al. 2014; Mirceta et al. 2013). Furthermore, manyexamples are found in the case of “parallel speciation” for species that are closephylogenetically (this is why the term parallel is used for different populations ofthe same species; Jones et al. 2012; Elmer and Meyer 2011; Soria-Carrasco et al.2014). It should be noted that in these species the genetic origin of the convergencecould be due to the standing variation or hybridization.

1C) Convergence at the character or the organism levels

The definition of convergence can be based on specific traits (see, for example,Pankey et al. 2014) or based on the total organism: global similarities based onseveral characters (see, for example, Losos et al. 2011; Malher et al. 2013).

1.2 Road Map

1.2.1 Re-analyses of Known Cases and Analyses of NewCases of Convergent Evolution at the Phenotypicand Environmental Level

Concerning these three points, our proposition is to re-analyse the described andnew cases on a phylogenetic basis and to extract on one side the cases corre-sponding to isoconvergent evolution and on the other those corresponding toalloconvergence.

1.2.1.1 Bibliographic Analysis and Research of Undescribed Cases

Research of known casesThis is the strategy we propose for the research of known cases. Research can bedone on the ISI database (http://www.webofknowledge.com/) using convergent*AND evolution* keywords. In order to complete this search, several books can be

1 Road Map to Study Convergent Evolution: A Proposition … 7

used: the McGhee (2011), the Conway Morris (2003) and the Sanderson andHufford (1996), but also the website dedicated to the convergent evolution andcoordinated by Conway Morris, i.e. map of life at http://www.mapoflife.org/, theWikipedia site at http://en.wikipedia.org/wiki/List_of_examples_of_convergent_evolution, as well as the compilation realized by Connie Barlow at www.thegreatstory.org/convergence.pdf. General reviews like that of Gordon and Notar(2015) will also be used.

Research of new casesFollowing the strategy previously proposed by Hiller et al. (2012), one can performa systematic search for all the characters of a phylogenetic group and find one orseveral characters that have a paraphylogenetic distribution; in other words, to mapall the characters on a phylogenetic tree and decipher the character that evolves inan independent manner. A database on phenotypes in the context of the phylogenyalready exists: the morphobank, and is available online at http://www.morphobank.org/ (O’Leary and Kaufman 2011).

Habitat shift convergenceThe search can be done by screening the ISI database using the key words con-vergent* AND environment* or, in an indirect manner, convergent* evolution andadaptation*. We can perform analysis showing the paraphyletic environmentaldistribution for a given clade, for example the aquatic way of life of mammals. Atthis stage, it is important to define the environment; an international effort is on theway (Buttigieg et al. 2013).

1.2.1.2 Definition of the Phenotype

Functional level, morphological level, distinction between discrete and con-tinuous characters. Character convergence/Global convergence.

We now consider the different levels of biological organization to which con-vergence concepts can be applied. At the physiological level, one can distinguishdifferent sublevels: biochemical (metabolism), electrophysiological, mechanical andphysical. It is important to add here another level: the one of the environmentaldemand and the possibility that different solutions can be found to face the sameproblem. One of the classical examples is predators that confront prey containingtoxic substances and may either evolve resistance or avoid eating the part of the bodythat contains the toxin. It should also be noted that very different morphologies mayproduce similar functional capabilities. For example, labrid fish with many differentjaw structures can produce the same suction force (Alfaro et al. 2005).

Further, a physiological character can be linked to several morphological traits.For example, talking about ultrasonic detection, the deformation of the cochlea isisoconvergent in the ultrasonic hearing lineage but bats have external ears thatallow amplifying the signal which is not the case for the dolphins. A given functionor physiology can also be the addition of several functions. The example of

8 P. Pontarotti and I. Hue

echolocation is a good one as it corresponds in fact to the addition of severaldifferent functions, and it is important to dissect the functions and then look if eachfunction (subfunction) is supported by one or several morphological characters.

Once the character is defined, we then have to classify the cases as isoconver-gence or alloconvergence. In the case of alloconvergence, because the characterunder investigation occurred in an independent manner, it can be used for statisticalanalysis as an independent observation and link, for example, to other parameterssuch as environment. However, the link with the other biological levels cannot bedone (see above). In the case of isoconvergent evolution, many tests can be done(see after); furthermore, the link with the other biological levels can be investigated,and this will be discussed in the following paragraph.

1.2.1.3 Isoconvergence Detection and Test

Methods have been developed for both discrete and continuous traits in the case ofisoconvergent evolution with well-defined characters.

Discrete Characters, Character per Character at the Phenotype Level

Identifying isoconvergence starts by an ancestral state reconstruction of the iso-convergent trait. For example, this method has provided support for isoconvergentevolution of plumage coloration in Icterus orioles (Omland and Lanyon 2000) andthe origin of photosphores in squids (Pankey et al. 2014). In such an analysis, thephenotype is reconstructed over the phylogeny, and independent origins are takenas an evidence of isoconvergence.

Continuous Trait at Phenotype Level

Most of the time this approach is used for multicharacter traits to test isoconvergentevolution at the global (species) level, but it can also be used for a simple charactertrait, biochemical activity, for example.

Detection of the isoconvergent evolutionA first approach has been developed by Muschick et al. (2012) that used a phenotypeapproach to test for convergence in cichlid fishes by considering that convergenceshould result in a pattern of reduced phenotypic differentiation when compared withphylogenetic distance. They calculated Euclidean distances between species for themorphological traits of interest and plotted them against the phylogenetic distances.They then used simulations to identify instances where phenotypic divergence wassignificantly lower than expected, based on phylogenetic distance.

A second approach was described by Ingram and Mahler (2013), whichexplicitly modelled trait evolution onto a phylogeny to identify convergent

1 Road Map to Study Convergent Evolution: A Proposition … 9

evolution. Their method (called SURFACE) takes a continuous trait and fitsOrnstein–Uhlenbeck models, with varying numbers of selective regimes and withshifts at varying points on the tree. Akaike’s information criterion is then used toselect the best fitting model. Convergence is identified by the independent adoptionof the same similar shift at multiple points on the phylogeny. This method repre-sents a technique to identify when convergence has occurred.

Test to evidence whether we have more isoconvergence than expectedIn order to quantify the strength of isoconvergence, Arbuckle et al. (2014) devel-oped an index (Wheatsheaf) that quantifies the strength of convergence as the ratioof the average pairwise distance between all the taxa in the dataset to the averagepairwise distance between all putatively convergent taxa. This kind of analysis isdone usually for a group of characters in order to test a global convergence (seeMalher et al. 2013; Moen et al. 2016 and for a review Stayton 2015).

1.2.2 Isoconvergence: Link Between the DifferentBiological Levels: From the Genotypeto the Phenotype

Because it is possible that isoconvergence at the phenotype level could be linked tosimilar mechanisms at other biological levels (multilevel isoconvergence), it isimportant to show here how the different biological levels are linked; this isdescribed in the following paragraph (Fig. 1.2).

Fig. 1.2 The different biological levels to be analysed

10 P. Pontarotti and I. Hue

1.2.2.1 Genetics/Epigenetics Mechanism Linking the Genotypeto the Phenotype

The causative mutation for a phenotypic shift should be found at the DNA level:point mutation, indel deletion, horizontal gene transfer, at the coding level ornon-coding level (repetitive element, non-coding RNA, micro RNA…). At thenon-coding level, the consequence of a mutation can be at epigenetic and tran-scriptional level (Lynch et al. 2015). To better analyse the consequence of themutation in the broader context of convergent evolution, we argue that it isimportant to have a network view instead of a gene-centred view (as genes do notwork in isolation).

The most integrative way to understand the role of the different genes andcorresponding networks of genes in development and in the relation “phenotype togenotype” is given by the work of Davidson and collaborators in seminal papers ongene regulatory network (GRN) and differentiation gene battery (DGB). The GRNshows a hierarchical organization. The GRNs establishing this initial postgastrularregulatory state, including the signalling interactions that help to establish domainboundaries, could be termed GRN level 1 (GRN1). The progenitor fields for thefuture adult body parts are later demarcated by signals plus local regulatory spatialinformation, and given regulatory states are established in each such field by theearliest body part-specific GRNs. Many such progenitor fields are thus set up duringpostgastrular embryogenesis. This corresponds to the second hierarchical level 2(GRN2). Each progenitor field is then divided up into the subparts that will togetherconstitute the body part, where the subdivisions are initially defined by installationof unique GRNs producing unique regulatory state GRN 3. The next-level termi-nation of the developmental process in each region of the late embryo, the GRNsspecifying the several individual cell types and deployed in each subpart of eachbody part, is the level 4. These GRNs control the expression of differentiation genebatteries, the final outputs of each cell type.

The evolution of forms or the apparition of a new structure is a developmentalprocess that can be due to a neo-expression of the GRN levels 1, 2, 3. The apparitionof a new cell type can be due to the neo-expression of GRN but at different levels 3and 4 as well as from the neo-expression of the differentiation gene battery (due tothe neo-expression of GRN or due to new promoter elements in cis position thatcould come from a retro-element, for example) (Lynch et al. 2015). As a conse-quence, the repeated variation should also be analysed at the transcriptional andepigenetic level (Pankey et al. 2014; Gallant et al. 2014; Pfenning et al. 2014).

Concerning physiological character (function), as said above, a new functioncould correspond to the set-up of complex organs. At the physiological level, wecan distinguish several sublevels: biochemical (metabolism), electrophysiological,mechanical and physical. For the biochemical and the electrophysiological levels,the function will be supported by the DGB even though the transcriptional state ofthe involved genes may be regulated at an upper level in the GRN (see, forexample, the electrogenesis case: Gallant et al. 2014; Zakon et al. 2006). In the caseof a simple biochemical shift, this can be explained by a mutation on a DGB (Hiller

1 Road Map to Study Convergent Evolution: A Proposition … 11

et al. 2012). This will be also the case for metabolic enzyme (Hiller et al. 2012) orfor a gene coding a protein that gives a resistance to a poison or anantibiotic-resistant enzyme. Note also that a shift in the activity could be due to adifference at transcriptional or post-transcriptional level in the case of biochemicalshift. It will thus be important to scan at the protein level but also at the tran-scriptional and epigenetic level.

1.2.2.2 Link Between the Different Biological Levels

It has to be noted that isoconvergence at one level could be correlated with either iso-or alloconvergence, or difference at the other biological levels; one of the paradig-matic examples is the flying capacity of bats and birds which came from an organthat allowed their common ancestor to walk. This is a clear case of isoconvergentevolution at the functional level, but in that case the organ involved in this functionevolved via divergent evolution from the same ancestral character. Another example,the fly for butterfly, birds and mammals, is a case of alloconvergent evolution, sinceat the functional level the ancestral structure was not involved in the walk in theinsect ancestor. The organ evolved via different arms in the case of birds/mammals,lateral lobes in the case of insects via divergent evolution. Another example can befound in the case of enzyme, the function of which evolved via alloconvergentevolution, with different ancestral functions and the same derived function.However, they evolved from a different structure in a divergent manner.

In the case of alloconvergent evolution, it is therefore difficult to make the linkbetween structure and function. For example, in the case of enzyme, we cannotpredict the involved amino acids at the sequence or biophysical levels. When theancestor is different, many ways are possible to obtain the derived function, exceptif alloconvergent evolution is also present at the structural level. Therefore, the linkbetween the evolution of the structure and the evolution of the function is notpossible in that case.

In the case of enzymatic isoconvergent evolution, where the process starts likelyfrom the same structure and the same function, it is plausible that the same changeoccurred at that protein level (amino acid sequence, biophysical properties). Severalreports show that for receptors or enzymes: Mirceta et al. (2013), Ujvari et al.(2015). This phenomenon could be due to epistasis and pleiotropy (Martin andOrgogozo 2013).

In the case of isoconvergence at functional level, the link with the other levelwill be less evident, as seen above it is at the superior biological level than at themorphological one. So if isoconvergent evolution is present at the physiologicallevel, the morphological level has to be checked (whenever possible) before pro-ceeding to the other levels (genetic, epigenetic). If one has access only at thephysiology, without knowledge of the supporting structure (e.g. adaptation toaquatic life), the relation with the other biological levels will be less evident.However, as described by many authors, the more similar the species, the higher thechance to have the same evolving mechanism (Conte et al. 2012).

12 P. Pontarotti and I. Hue

1.2.2.3 Origin of the Genetic Variation: De Novo Versus StandingVariation

In the case of iso- or alloconvergent evolution at the nucleic acid and amino acidlevels, the convergence could be due to independent mutations in the differentlineages (de novo mutations) or due to standing variation or any other processescorresponding to horizontal gene transfer. This has been explained in the literatureby Stern (2013), or Martin and Orgogozo (2013). The question is how to distin-guish standing variation from de novo mutations. This point is important since,depending on the origin of the mutations, we cannot use the same model; in the caseof de novo mutation, we need to use a model based on the phylogeny, whereas inthe case of transfer or substitutions coming from the standing variation, differentmodels are needed.

In general, the longer the time of separation between species where isoconver-gence is found ancient, the weaker the chances to find pre-existing mutationsbetween such species. Several mechanisms could explain why a pre-existingmutation will be shared across “distant” species: (i) incomplete lineage sorting due torapid speciation, (ii) balancing selection, (iii) hybridization, (iv) paleo-hybridizationin a group that separated but entered in contact again, followed by a new separationor any case of horizontal gene transfer (Bird et al. 2015). The different trait includedat the genome level will not follow a bifurcative story, even if thepaleo-hybridization event is old. In the case of paleo-hybridization, the conflictingtopology will occur at the time the events occurred and a little bit after that, due to theallele sorting. For example, there is a probability that hybridization occurred at thebase of the mammalian radiation, and this is why we may have conflicting historiesbetween the mammalian lineages (Hallström and Janke 2010).

1.2.2.4 De Novo Isoconvergent Evolution from the Genotypeto the Phenotype

Once the isoconvergent evolution has been identified at the phenotype level, thehypothesis of co-convergent evolution with the other levels can be tested (seeFigs. 1.2 and 1.3). To do so, one needs to perform ancestral reconstruction andshow that the apomorphy occurred in a convergent manner. This should be done foreach biological level. In cases where the events at the different levels are con-comitant (co-convergence), a possible cause-to-effect relationship across the levelswill be evidenced (Fig. 1.2 and 1.3).

Direct link from genome to phenotypeAmino acid substitution that could cause a character shift (Fig. 1.3).

Zhang and Kumar (1997) as well as Foote et al. (2015) linked the convergentevolution at the amino acid level with the convergent evolution at the phenotypiclevel. The probable evolution of the amino acid is evaluated using a likelihood-basedmodel. The next step is then to detect the shift from the plesiomorphic to the

1 Road Map to Study Convergent Evolution: A Proposition … 13

isoconvergent apomorphic sites. This is done by reading automatically the phylo-genetic tree. However, in these studies and in all the studies published so far thatlinked the phenotype to the genotype, the reconstruction of the phenotypic characterevolution (most of time discrete characters) is performed with an algorithm based onthe Mirkin et al.’s (2003) approach, even if the authors do not write it explicitly. Thisalgorithm evidences where the characters are present on a group of leaves and wherethey are absent, to conclude that the characters appeared after the separation of thegroups having or not the characters, which is an approximation, except if one hasaccess to fossils or other information as in the case of the marine mammals whereone can guess that the ancestral reconstruction of the way of life is correct since lotsof information show that the ancestors of the marine groups were terrestrial.

It should be noted that in general, character reconstruction could be done usingstochastic approaches (maximum likelihood; for example, see for review RoyerCarenzi et al. 2013).

Fig. 1.3 Schematic example of how to link the phenotype to the other biological levels on aphylogenetic tree (here the link between a mutation in a protein and a given function). Step 1:Phenotype history reconstruction. Step 2: Map of the phenotypic shifts in our phylogenetic tree.Step 3: Proteome history reconstruction. Step 4: Map of all the convergent (allo or iso)substitutions in our phylogenetic tree, selecting in yellow the ones that co-occur with thephenotypic shift. Step 5: Ortholog genes having convergent substitutions are candidates for thephenotype shift. This ranking can then be redefined with other predictive analysis such as functionprediction (see Parker et al. 2013; Foote et al. 2015). Step 6: Experimental task

14 P. Pontarotti and I. Hue

Of course, convergence at amino acid levels due to noise is possible and gives afalse signal. The significance increases if many positions are needed to get the newphenotype (and if several genes are needed for a given phenotype). However, it hasto be noted that only few amino acid changes can lead to a shift of the function for agiven protein and that, even if the statistical test is inconclusive, the changes can beinvolved in the phenotypic shift.

Zhang and Kumar (1997) developed a test where they looked whether the numberof alloconvergent or isoconvergent amino acids that occurred in two or moreorthologues is above what is expected. A similar but different test has been devel-oped by Castoe et al. (2009). In their analysis, these authors were interested to testwhether two species evolved more convergently (at least) for a part of their genomes.The authors tested all the branches and tested whether the chosen species evolvedmore convergently than the other couple of species or against simulated data.

Detection of evolutionary isoconvergent “characteristics” at genomic levelCoding sequence

If the isoconvergent evolution at the phenotypic level results from adaptation, itis possible that the causal mutation at the genome level has been selected, and a testto detect positive selection can be used so that the protein that evolved underpositive selection in a convergent manner can therefore be detected. The nice thingwith this method is that even if different amino acids are involved in a similarfunctional shift in the coding sequence, the genotype to the phenotype link ispossible (this approach has been used by Parker et al. (2013), Foote et al. (2015),for review: Levasseur et al. 2007). Concerning the shift on the constraints seen oncommon or different sites of the same protein (Gu et al. 2001; Lichtage et al. 1996;Gribaldo et al. 2003), a general relaxation of a protein (different sites on the sameprotein) could mean that some of its functions are less important for the fitness ofthe species. Here again, we can test whether we have more relaxation than expected.An extreme case is the loss of function (pseudogenization). This has been nicelyshown in the case of the convergent loss of the GULO gene involved in the vitaminC synthesis (Hiller et al. 2012).

We could also have the case of different genes involved in the same convergentphenotype but belonging to the same network, for example: all the orthologues of anetwork where at least one gene evolves under positive selection or has the sameconstraint in the different species presenting the convergent phenotypic character.The protein corresponding to these genes could be involved in the same cascade,the same network, the same pathway (Foote et al. 2015).

Non-coding sequenceIn analyses at the promoter level, one first needs to identify the promoters. This ispossible for species that are close phylogenetically as promoter conservation ispossible. In any cases, it is easier to detect convergent evolution in the case of aderegulation (loss of the expression in a given tissue, for example). In that case, thepromoters should evolve quicker than expected, as reported for Shaven Baby

1 Road Map to Study Convergent Evolution: A Proposition … 15

(Frankel et al. 2012). Therefore, the cause of an expression loss should be possibleto detect.

However, it would be really difficult to evidence the causal mutation of a newexpression territory (especially if the mutation is not the same). Nonetheless, manyexamples in the literature show that a new expression territory could be due to theintegration of a retro-element in the gene promoter (Lynch et al. 2015) or a newfunction could be due to an LTR coding sequence (Naville et al. 2016; Pavlicevet al. 2015), and one can look for such sequences surrounding the orthologousgenes in the species that show convergent evolution at phenotypic level and evi-dence their absence in the species where the phenotype is absent.

Molecular level: convergent evolution of physical–chemical proteincharacteristicsA new protein function could not be due to specific amino acid changes but due tothe overall physical properties. In that case, it is possible to perform an ancestralreconstruction of the physical properties of the protein families and looked forisoconvergent shift and then looked for the co-isoconvergence for the studiedfunction (see Mirceta et al. 2013; Ujvari et al. 2015).

Detection of genetic events other than substitutionsOther genetic events can be evidenced (Gouret et al. 2011; Dainat et al. 2012;Dainat and Pontarotti 2014; Le et al. 2012; Paganini et al. 2012) and linked to thephenotype shift (Cayrou et al. 2012; Levasseur et al. 2012).

1.2.2.5 Detection of Isoconvergent Evolution at IntermediateBiological Levels

Expression dataThe difference is observed directly in one gene of the GRN. As said above, it willbe really difficult to identify the causative mutation. Therefore, in that case, oneneeds to investigate the expression data of the concerned tissues. Once the geneco-opted is evidenced, one can look at the cis regulatory region or search for anepigenetic signature such as the methylation status of the gene that has been shownto be over- or underexpressed as reported by Lynch et al. (2015).

We therefore need to have access to the concerned tissues and perform expressionstudies using a methodology similar to the one described by Pankey et al. toevidence the similarity between the two isoconvergent derived tissues and thedifference with the others. These analyses can then be integrated in an interactionnetwork, signalling pathway, metabolic pathway or gene regulation network(Gallant et al. 2014; Pankey et al. 2014; Pfenning et al. 2014).

It has to be noted that in that case, the authors did not use ancestral reconstructionfollowed by the search of the shift. They performed a comparative transcriptomicanalysis between the new cells or organs, as compared to the other cells or organs of

16 P. Pontarotti and I. Hue

the species where the convergent cell tissues have been evidenced, and looked forwhat is common for the isoconvergent tissues and different for the others.

Integration of the different biological levelsThe further step is to integrate the different levels seen in Fig. 1.1, in order toexplain the cause/consequence effect from the genotype to the phenotype. To thebest of our knowledge, only one example of co-convergence at sequence, expres-sion and phenotype has only been done in one case where mutation in the promoterwas linked to the expression loss, leading to the loss of a phenotype (see for acomplete story Stern and Frankel 2013). In that case, the promoters should evolvequicker than expected. This is what is happening for Shaven Baby (Frankel et al.2012), a GRN gene that controls several differential battery genes. In that case, theloss of a phenotypic trait should be possible to detect. However, to evidence thecausal mutation of a new expression territory would be really difficult. However, asmany examples in the literature show that a new expression territory could be due toa new integration of a retro-element in the gene promoter (Lynch et al. 2015), theidentification of similar retro-element in front of differential expressed gene shouldbe possible.

1.2.3 Database Organization

Based on the points discussed in this article, it will be useful to develop a databaseand mandatory that this database uses an ontology that takes into account thedifferent levels and links one to the other (Fig. 1.2). The database could be filled inby the different collaborators. The database we propose will offer a new way toanalyse the different types of convergent evolution: separate apparition of the samederived state (apomorphic) in at least two distinct lineages from the same ancestralstate (isoconvergence) or from different ancestral states (alloconvergence), con-sidering the different biological levels. In other words, an amino acid substitutioncould induce a shift in an enzymatic function. The latter could be linked to a shift inthe metabolite that can infer a shift in the phenotype at morphological and/or atphysiological levels. Thus, because of the complexity of the reality and the widevariety of the data, this database will be flexible enough to store biological data aswell as to incorporate different biological levels.

1.3 Conclusion/Road Map

In conclusion, we propose (1) to re-analyse all the cases of convergent evolution atthe phenotype level described in the literature and sort out cases of isoconvergence,(2) to identify undescribed cases of isoconvergent evolution, (3) to use the strategies

1 Road Map to Study Convergent Evolution: A Proposition … 17

developed in this article to study these cases and on selected school case studies as,for example, the evolution of the active electro-localization in fishes and (4) tocreate an international effort by the access of the database that is able to integratethe different cases of isoconvergence at different biological levels.

Acknowledgments We would like to acknowledge our colleagues from the EBM team, as well asBenoit Heulin, Olivier Sandra and Laurent Journot for helpful discussions and Christophe Kloppand Mike Speed for critical reading of the manuscript.

References

Alderson RG, Barker D, Mitchell JB (2014) One origin for metallo-β-lactamase activity, or two?An investigation assessing a diverse set of reconstructed ancestral sequences based on a sampleof phylogenetic trees. J Mol Evol 79(3–4):117–129

Alfaro ME, Bolnick DI, Wainwright PC (2005) Evolutionary consequences of many-to-onemapping of jaw morphology to mechanics in labrid fishes. Am Nat 165:E140–E154

Arbuckle K, Bennett CM, Speed MP (2014) A simple measure of the strength of convergentevolution. Methods Ecol Evol 5(7):685–693

Bird DM, Jones JT, Opperman CH, Kikuchi T, Danchin EG (2015) Signatures of adaptation toplant parasitism in nematode genomes. Parasitology 142(Suppl 1):S71–S84

Buttigieg PL, Morrison N, Smith B, Mungall CJ, Lewis SE, ENVO Consortium (2013) Theenvironment ontology: contextualising biological and biomedical entities. J Biomed Semantics11;4(1):43

Conway Morris S eds (2003) Life’s solution: inevitable humans in a lonely universe. CambridgeUniversity Press, Cambridge

Castoe TA, de Koning AP, Kim HM, Gu W, Noonan BP, Naylor G, Jiang ZJ, Parkinson CL,Pollock DD (2009) Evidence for an ancient adaptive episode of convergent molecularevolution. Proc Natl Acad Sci USA 106(22):8986–8991

Cayrou C, Henrissat B, Gouret P, Pontarotti P, Drancourt M (2012) Peptidoglycan: a post-genomicanalysis. BMC Microbiol 12:294

Conte GL, Arnegard ME, Peichel CL, Schluter D (2012) The probability of genetic parallelism andconvergence in natural populations. Proc Biol Sci 279(1749):5039–5047

Dainat J, Pontarotti P (2014) Methods to study the occurrence and the evolution of pseudogenesthrough a phylogenetic approach. Methods Mol Biol 116:87–99

Dainat J, Paganini J, Pontarotti P, Gouret P (2012) GLADX: an automated approach to analyzethe lineage-specific loss and pseudogenization of genes. PLoS One 7(6)

Dick R, Rattei T, Haslbeck M, Schwab W, Gierl A, Frey M (2012) Comparative analysis ofbenzoxazinoid biosynthesis in monocots and dicots: independent recruitment of stabilizationand activation functions. Plant Cell 24(3):915–928

Doolittle RF (1994) Convergent evolution: the need to be explicit. Trends Biochem Sci 19:15–18Elmer KR, Meyer A (2011) Adaptation in the age of ecological genomics: insights from

parallelism and convergence. Trends Ecol Evol 26(6):298–306Foote AD, Liu Y, Thomas GW, Vinař T, Alföldi J, Deng J, Dugan S, van Elk CE, Hunter ME,

Joshi V, Khan Z, Kovar C, Lee SL, Lindblad-Toh K, Mancia A, Nielsen R, Qin X, Qu J,Raney BJ, Vijay N, Wolf JB, Hahn MW, Muzny DM, Worley KC, Gilbert MT, Gibbs RA(2015) Convergent evolution of the genomes of marine mammals. Nat Genet 47(3):272–275

Frankel N, Wang S, Stern DL (2012) Conserved regulatory architecture underlies parallel geneticchanges and convergent phenotypic evolution. Proc Natl Acad Sci USA 51:20975–20979

18 P. Pontarotti and I. Hue

Gallant JR, Traeger LL, Volkening JD, Moffett H, Chen PH, Novina CD, Phillips GN Jr, Anand R,Wells GB, Pinch M, Güth R, Unguez GA, Albert JS, Zakon HH, Samanta MP, Sussman MR(2014) Genomic basis for the convergent evolution of electric organs. Science 344(6191):1522–1525

Gherardini PF, Wass MN, Helmer-Citterich M, Sternberg MJE (2007) Convergent evolution ofenzyme active sites is not a rare phenomenon. J Mol Biol 372(3):817–845

Gordon MS, Notar JC (2015) Can systems biology help to separate evolutionary analogies(convergent homoplasies) from homologies? Prog Biophys Mol Biol 117(1):19–29

Gouret P, Paganini J, Dainat J, Louati D, Darbo E, Pontarotti P, Levasseur A (2011) Integration ofevolutionary biology concepts for functional annotation and automation of complex research inevolution: the multi-agent software system. In: P Pontarotti (ed) Evolutionary biology-conceptsbiodiversity, macroevolution and genome evolution, pp 71–87

Gribaldo S, Casane D, Lopez P, Philippe H (2003) Functional divergence prediction fromevolutionary analysis: a case study of vertebrate haemoglobin. Mol Biol Evol 11:1754–1759

Gu X (2001) Maximum likelihood approach for gene family evolution under functionaldivergence. Mol Biol Evol 18:453–464

Hallström BM, Janke A (2010) Mammalian evolution may not be strictly bifurcating. Mol BiolEvol 27(12):2804–2816

Hiller M, Schaar BT, Indjeian VB, Kingsley DM, Hagey LR, Bejerano GA (2012) A “forwardgenomics” approach links genotype to phenotype using independent phenotypic losses amongrelated species. Cell Rep 2(4):81723

Ingram T, Mahler DL (2013) SURFACE «A simple measure of the strength of convergentevolution»: detecting convergent evolution from comparative data by fitting Ornstein-Uhlenbeck models with stepwise Akaike Information Criterion. Methods Ecol Evol 4:416–425

Jones FC, Grabherr MG, Chan YF, Russell P, Mauceli E, Johnson J, Swofford R, Pirun M,Zody MC, White S, Birney E, Searle S, Schmutz J, Grimwood J, Dickson MC, Myers RM,Miller CT, Summers BR, Knecht AK, Brady SD, Zhang H, Pollen AA, Howes T, Amemiya C,Broad Institute Genome SequencingPlatform, Whole Genome Assembly Team, Baldwin J,Bloom T, Jaffe DB, Nicol R, Wilkinson J, Lander ES, Di Palma F, Lindblad-Toh K,Kingsley DM (2012) The genomic basis of adaptive evolution in threespine sticklebacks.Nature 484(7392):55–61

Kopp A (2009) Metamodels and phylogenetic replication: a systematic approach to the evolutionof developmental pathways. Evolution 63(11):2771–2789

Le PT, Ramulu HG, Guijarro L, Paganini J, Gouret P, Chabrol O, Raoult D, Pontarotti P (2012)An automated approach for the identification of horizontal gene transfers from completegenomes reveals the rhizome of Rickettsiales. BMC Evol Biol 12:243

Levasseur A, Orlando L, Bailly X, Milinkovitch MC, Danchin EG, Pontarotti P (2007) Conceptualbases for quantifying the role of the environment on gene evolution: the participation ofpositive selection and neutral evolution. Biol Rev Camb Philos Soc 82(4):551–572

Levasseur A, Paganini J, Dainat J, Thompson JD, Poch O, Pontarotti P, Gouret P (2012) Thechordate proteome data base. Evol Bioinform Online 8:437–447

Lichtarge O, Bourne HR, Cohen FE (1996) An evolutionary trace method defines binding surfacescommon to protein families. J Mol Biol 257(2):342–358

Losos JB (2011) Convergence, adaptation, and constraint. Evolution 65(7):1827–1840Lynch VJ, Nnamani MC, Kapusta A, Brayer K, Plaza SL, Mazur EC, Emera D, Sheikh SZ,

Grützner F, Bauersachs S, Graf A, Young SL, Lieb JD, DeMayo FJ, Feschotte C, Wagner GP(2015) Ancient transposable elements transformed the uterine regulatory landscape andtranscriptome during the evolution of mammalian pregnancy. Cell Rep 10(4):551–561

Mahler DL, Ingram T, Revell LJ, Losos JB (2013) Exceptional convergence on the macroevo-lutionary landscape in island lizard radiations. Science 341(6143):292–295

Martin A, Orgogozo V (2013) The Loci of repeated evolution: a catalog of genetic hotspots ofphenotypic variation. Evolution 67(5):1235–1250

McGhee GR (2011) Convergent evolution: limited forms most beautiful. The MIT Press,Cambridge

1 Road Map to Study Convergent Evolution: A Proposition … 19

McGowen MR, Gatesy J, Wildman DE (2014) Molecular evolution tracks macroevolutionarytransitions in Cetacea. Trends Ecol Evol 29(6):336–346

Mirceta S, Signore AV, Burns JM, Cossins AR, Campbell KL, Berenbrink M (2013) Evolution ofmammalian diving capacity traced by myoglobin net surface charge. Science 340:1234192

Mirkin BG, Fenner TI, Galperin MY, Koonin EV (2003) Algorithms for computing parsimoniousevolutionary scenarios for genome evolution, the last universal common ancestor anddominance of horizontal gene transfer in the evolution of prokaryotes. BMC Evol Biol 3:2

Moen DS, Morlon H, Wiens JJ (2016) Testing convergence versus history: convergence dominatesphenotypic evolution for over 150 million years in frogs. Syst Biol 65(1):146–160

Muschick M, Indermaur A, Salzburger W (2012) Convergent evolution within an adaptiveradiation of cichlid fishes. Curr Biol 22(24):2362–2368

Naville M, Warren IA, Haftek-Terreau Z, Chalopin D, Brunet F, Levin P, Galiana D, Volff JN(2016) Not so bad after all: retroviruses and LTR retrotransposons as a source of new genes invertebrates. Clin Microbiol Infect. doi:10.1016/j.cmi.2016.02.001. (Epub ahead of print)

O’Leary MA, Kaufman S (2011) MorphoBank: phylophenomics in the “cloud”. Cladistics27:529–537

Omland KE, Lanyon SM (2000) Reconstructing plumage evolution in orioles (Icterus): repeatedconvergence and reversal in patterns. Evolution 54(6):2119–2133

Paganini J, Campan-Fournier A, Da Rocha M, Gouret P, Pontarotti P, Wajnberg E, Abad P,Danchin EGJ (2012) Contribution of lateral gene transfers to the genome composition andparasitic ability of root-knot nematodes. Plos One 7(11):e50875

Pankey MS, Minin VN, Imholte GC, Suchard MA, Oakley TH (2014) Predictable transcriptomeevolution in the convergent and complex bioluminescent organs of squid. Proc Natl AcadSci USA 11(44):E473642

Parker J, Tsagkogeorga G, Cotton JA, Liu Y, Provero P, Stupka E, Rossiter SJ (2013)Genome-wide signatures of convergent evolution in echolocating mammals. Nature 502(7470):228–231

Pavlicev M, Hiratsuka K, Swaggart KA, Dunn C, Muglia L (2015) Detecting endogenousretrovirus-driven tissue-specific gene transcription. Genome Biol Evol 7(4):1082–1097

Pfenning AR, Hara E, Whitney O, Rivas MV, Wang R, Roulhac PL, Howard JT, Wirthlin M,Lovell PV, Ganapathy G, Mouncastle J, Moseley MA, Thompson JW, Soderblom EJ, Iriki A,Kato M, Gilbert MT, Zhang G, Bakken T, Bongaarts A, Bernard A, Lein E, Mello CV,Hartemink AJ, Jarvis ED (2014) Convergent transcriptional specializations in the brains ofhumans and song-learning birds. Science 346(6215):1256846

Rosenblum EB, Parent CE, Brandt EE (2014) The molecular basis of phenotypic convergenceannual review of ecology. Evol Syst 45:203–226

Royer-Carenzi M, Pontarotti P, Didier G (2013) Choosing the best ancestral character statereconstruction method. Math Biosci 242(1):95–109

Sanderson MJ, Hufford L (eds) (1996) Homoplasy: the recurrence of similarity in evolution.Academic Press, New York

Soria-Carrasco V, Gompert Z, Comeault AA, Farkas TE, Parchman TL, Johnston JS, Buerkle CA,Feder JL, Bast J, Schwander T, Egan SP, Crespi BJ, Nosil P (2014) Stick insect genomesreveal natural selection’s role in parallel speciation. Science 344(6185):738–742

Stayton CT (2015) The definition, recognition, and interpretation of convergent evolution, and twonew measures for quantifying and assessing the significance of convergence. Evolution269(8):2140–2153

Stern DL, Orgogozo V (2008) The loci of evolution: how predictable is genetic evolution?Evolution 62(9):2155–2177

Stern DL (2013) The genetic causes of convergent evolution. Nat Rev Genet 14(11):751–764Stern DL, Frankel N (2013) The structure and evolution of cis-regulatory regions: the shaven baby

story Philos Trans R Soc. Lond B Biol Sci 68(1632):20130028Ujvari B, Casewell NR, Sunagar K, Arbuckle K, Wüster W, Lo N, O’Meally D, Beckmann C,

King GF, Deplazes E, Madsen T (2015) Widespread convergence in toxin resistance bypredictable molecular evolution. Proc Natl Acad Sci USA 112(38):11911–11916

20 P. Pontarotti and I. Hue

Zakon HH, Lu Y, Zwickl DJ, Hillis DM (2006) Sodium channel genes and the evolution ofdiversity in communication signals of electric fishes: convergent molecular evolution. PNAS103:3675–3680

Zhang J, Kumar S (1997) Detection of convergent and parallel evolution at the amino acidsequence level. Mol Biol Evol 14(5):527–536

1 Road Map to Study Convergent Evolution: A Proposition … 21

Chapter 2Analysing Convergent Evolution:A Practical Guide to Methods

Kevin Arbuckle and Michael P. Speed

Abstract Convergent evolution, or the independent evolution of similar traits, haslong been investigated and recognised as an important area of research for evolu-tionary biology. However, as with many areas of comparative biology, new phy-logenetic methods that enhance our ability to study convergence have arisen withgreater frequency in recent years. Consequently, we now have a wide range of toolsat our disposal and a rapidly developing conceptual framework to guide us in ouranalyses. This chapter aims to provide a practical guide for those interested inconvergent evolution that will enable new entrants to the field to quickly develop awell-rounded research agenda. Although some methods can be performed in otherpieces of (stand-alone) software, this guide will focus on the R statisticalenvironment.

2.1 Introduction

Convergent evolution is a common phenomenon across the diversity of livingorganisms. In essence, it refers to the independent evolution of some kind ofsimilarity between two or more organisms, as opposed to any similarity which is aresult of inheritance from a common ancestor. Convergent traits may be manifestacross a number of levels of biology including both function and form (Losos 2011;Speed and Arbuckle 2016). Convergence can be seen for example in many forms ofbehaviour, morphology and physiology (McGhee 2011), and in the structure andaction of molecules such as toxins or enzymes (Doolittle 1994). For instance, wecould consider the phenomenon of mimicry, in which one organism (perhaps aharmless viceroy butterfly, Limenitis archippus) evolves to appear like a different

K. Arbuckle (&) � M.P. SpeedDepartment of Ecology Evolution and Behaviour Biosciences Building,University of Liverpool, Crown Street, Liverpool L69 7ZB, UKe-mail: k.arbuckle@liverpool.ac.uk

M.P. Speede-mail: speedm@liverpool.ac.uk

© Springer International Publishing Switzerland 2016P. Pontarotti (ed.), Evolutionary Biology, DOI 10.1007/978-3-319-41324-2_2

23