Abstract

Recent molecular studies investigating higher-level phylogenetics of coleoid cephalopods (octopuses, squids and cuttlefishes) have produced conflicting results. A wide range of sequence alignment and analysis methods are used in cephalopod phylogenetic studies. The present study investigated the effect of commonly used alignment and analysis methods on higher-level cephalopod phylogenetics. Two sequence homology methods: (1) eye alignment, (2) implied alignment, and three analysis methods: (1) parsimony, (2) maximum likelihood, (3) Bayesian methodologies, were employed on the longest sequence dataset available for the coleoid cephalopods, comprising three mitochondrial and six nuclear loci. The data were also tested for base composition heterogeneity, which was detected in three genes and resolved using RY coding. The Octopoda, Argonautoidea, Oegopsida and Ommastrephidae are monophyletic in the phylogenies resulting from each of the alignment and analysis combinations. Furthermore, the Bathyteuthidae are the sister taxon of the Oegopsida in each case. However many relationships within the Coleoidea differed depending upon the alignment and analysis method used. This study demonstrates how differences in alignment and analysis methods commonly used in cephalopod phylogenetics can lead to different, but often highly supported, relationships.

INTRODUCTION

The class Cephalopoda comprises two extant subclasses, Nautiloidea (Nautilus and Allonautilus) and the Coleoidea. The Coleoidea contains two subdivisions, the Belemnoidea, which became extinct at the end of the Cretaceous, and the Neocoleoidea, which contains the octopuses, squids and cuttlefishes. Neocoleoid cephalopods are characterized by the reduction and internalization, or complete loss, of the shell and, as a result, they very rarely fossilize well. Therefore, very little information regarding the origins and relationships of extant coleoid cephalopods can be gleaned from the fossil record (Nishiguchi & Mapes, 2007).

Morphological studies have proved to be useful in classifying species within subfamilies and/or genera (e.g. Berthold & Engeser, 1987; Clarke, 1988; Khromov, 1990; Voight, 1993a,b; Young & Vecchione, 1996; Voss, 1988), but less so in determining higher-level relationships. Morphological studies attempting to resolve these relationships have been constrained by the number of characters used with confidence due to “problems primarily involving character independence, apomorphic ‘loss’, or assessment of homology/homoplasy” (Young & Vecchione, 1996).

Presently, extant coleoids are divided into two superorders, Decapodiformes and Octopodiformes (Berthold & Engeser, 1987). In his website ‘The Fossil Coleoidea Page’ (http://userpage.fu-berlin.de/~palaeont/fossilcoleoidea/welcome.html), Engeser draws attention to the fact that the term Octopodiformes is in use elsewhere and suggests the use of Vampyropoda (Boletzky, 1992) instead. The Decapodiformes (Decembrachiata Winckworth sensu Engeser, loc. cit.) contains the orders Teuthoidea [suborders Myopsida (closed-eye squids) and Oegopsida (open-eye squids) and Sepioidea (families Idiosepiidae (pygmy squid), Sepiidae (cuttlefishes), Spirulidae (ram's horn squid), Sepiolidae (bobtail squids) and Sepiadariidae (bottletail squids)]. Current debate exists on the validity of the ordinal level of classification (Naef, 1921–1923; Voss, 1977; Berthold & Engeser, 1987; Young & Vecchione, 1996). Furthermore, Lindgren, Giribet & Nishiguchi (2004) question whether the suborder Oegopsida is monophyletic.

The Octopodiformes contains the orders Vampyromorpha (vampire ‘squid’) and Octopoda (pelagic and benthic octopuses), hence the name Vampyropoda (Boletzky, 1992). A sister-taxon relationship between these two orders is accepted primarily based on morphology (Pickford, 1939; Boletzky, 1992; Young & Vecchione, 1996; Engeser, 1997; Young, Vecchione & Donovan, 1998; Carlini, Reece & Graves, 2000), but combined analysis using molecular and morphological data suggests a sister-taxon relationship between the Decapodiformes and Vampyromorpha (Lindgren et al., 2004). The Octopoda comprises the suborders Cirrata (deep-sea finned octopuses) and Incirrata (benthic octopuses and pelagic octopuses, including the argonautoids and blanket octopuses. A sister-taxon relationship between these suborders is also widely accepted (Grimpe, 1921; Naef, 1921–1923; Young & Vecchione, 1996; Voight, 1997). Phylogenetic relationships between the nine Incirrata families remain unresolved and have been debated in the literature (Naef, 1921–1923; Robson, 1929, 1931; Voss, 1977; Young & Vecchione, 1996; Voight, 1997).

In the mid 1990s the first studies using DNA sequence data to estimate phylogenetic relationships within cephalopods were reported (Bonnaud, Boucher-Rodoni & Monnerot, 1994, 1996, 1997; Boucher-Rodoni & Bonnaud, 1996). These studies sequenced portions of 16S rDNA, COII and COIII from 8 to 28 cephalopod taxa. These authors aligned their sequences by eye (with the aid of the secondary structure where possible) and analysed the data using neighbour-joining (NJ) and parsimony methods. Although the genes proved useful in helping resolve intrafamilial relationships, little resolution of higher-level relationships was recovered. Subsequently, molecular studies investigating higher-level phylogenetic relationships of cephalopods have sequenced additional mitochondrial genes (Carlini & Graves, 1999; Piertney et al., 2003; Nishiguchi, Lopez & Boletzky, 2004; Zheng et al., 2004; Guzik et al., 2005) including whole mitochondrial genomes (Yokobori et al., 2004; Akasaki et al., 2006) and also nuclear genes (Carlini et al., 2000; Warnke et al., 2003; Strugnell et al., 2004; Guzik et al., 2005; Strugnell et al., 2005) often from a greater number of taxa (Carlini & Graves, 1999; Anderson, 2000a, b; Carlini et al., 2000; Lindgren et al., 2004; Strugnell et al., 2005) (Table 1).

Table 1.

Summary of studies of the molecular phylogenetics of coleoid cephalopods.

Reference Focal taxa Genes used No. of species Sequence alignment method Analysis method(s) 
Bonnaud et al. (1994) Decapodiformes 16S 28 eye (2° structure) NJ, P 
Bonnaud et al. (1996) Decapodiformes 16S, COIII eye NJ, P 
Boucher-Rodoni & Bonnaud (1996)* Coleoidea 16S 10  NJ, P 
Bonnaud et al. (1997) Coleoidea COIII, COII 17 eye NJ, P 
Bonnaud et al. (1998) Onychoteuthidae 16 14 eye NJ, P 
Carlini & Graves (1999) Coleoidea COI 48 eye 
Anderson (2000) Loliginidae 16S, COI ∼30 Clustal and eye P, ML, LogDet 
Anderson (2000)* Loliginidae 16S, COI 53 Clustal and eye P, ML 
Carlini et al. (2000) Coleoidea actin 44 eye P, ML 
Carlini et al. (2001) Octopoda COI 29 eye P, ML 
Allcock & Piertney (2002) Octopodidae 16S Clustal X and eye NJ, P, ML 
Piertney et al. (2003) Cirrata 16S 27 Clustal X and eye NJ, P, ML 
Warnke et al. (2003) Decapodiformes complete 18S Clustal V, MegAlign, checked by eye NJ, P, ML 
Bonnaud et al. (2004) Nautilus complete 18S eye 2° structure 
Lindgren et al. (2004)Coleoidea complete 18S, 28S, hist. COI 60 POY 
Nishiguchi et al. (2004) Sepiolidae 12S, 16S, COI, 28S 30 POY 
Strugnell et al. (2004) Octopodiformes 16S, 12S, COI, rhod, pax-6, ODH  eye ML, Bayesian 
Yokobori et al. (2004) Coleoidea whole mitochondrial genome ClustalX NJ, ML 
Zheng et al. (2004) Decapodiformes COI, 16S 13 ClustalX v1.8 NJ, P 
Guzik et al. (2005) Octopodinae COIII, cyt b, ef-1α 30 Sequencher 3.1 P, ML, Bayesian 
Lindgren et al. (2005) Gonatidae 12S, 16S, COI 39 POY 
Strugnell et al. (2005) Coleoidea 16S, 12S, COI, rhod, pax-6, ODH 35 eye Bayesian, LogDet 
Takumiya et al. (2005) Coleoidea 12S, 16S, COI 36 SeqPup v. 0.9, ClustalX ver1.83 NJ, P, ML 
Akasaki et al. (2006) Coleoidea whole mitochondrial genome – ML 
Reference Focal taxa Genes used No. of species Sequence alignment method Analysis method(s) 
Bonnaud et al. (1994) Decapodiformes 16S 28 eye (2° structure) NJ, P 
Bonnaud et al. (1996) Decapodiformes 16S, COIII eye NJ, P 
Boucher-Rodoni & Bonnaud (1996)* Coleoidea 16S 10  NJ, P 
Bonnaud et al. (1997) Coleoidea COIII, COII 17 eye NJ, P 
Bonnaud et al. (1998) Onychoteuthidae 16 14 eye NJ, P 
Carlini & Graves (1999) Coleoidea COI 48 eye 
Anderson (2000) Loliginidae 16S, COI ∼30 Clustal and eye P, ML, LogDet 
Anderson (2000)* Loliginidae 16S, COI 53 Clustal and eye P, ML 
Carlini et al. (2000) Coleoidea actin 44 eye P, ML 
Carlini et al. (2001) Octopoda COI 29 eye P, ML 
Allcock & Piertney (2002) Octopodidae 16S Clustal X and eye NJ, P, ML 
Piertney et al. (2003) Cirrata 16S 27 Clustal X and eye NJ, P, ML 
Warnke et al. (2003) Decapodiformes complete 18S Clustal V, MegAlign, checked by eye NJ, P, ML 
Bonnaud et al. (2004) Nautilus complete 18S eye 2° structure 
Lindgren et al. (2004)Coleoidea complete 18S, 28S, hist. COI 60 POY 
Nishiguchi et al. (2004) Sepiolidae 12S, 16S, COI, 28S 30 POY 
Strugnell et al. (2004) Octopodiformes 16S, 12S, COI, rhod, pax-6, ODH  eye ML, Bayesian 
Yokobori et al. (2004) Coleoidea whole mitochondrial genome ClustalX NJ, ML 
Zheng et al. (2004) Decapodiformes COI, 16S 13 ClustalX v1.8 NJ, P 
Guzik et al. (2005) Octopodinae COIII, cyt b, ef-1α 30 Sequencher 3.1 P, ML, Bayesian 
Lindgren et al. (2005) Gonatidae 12S, 16S, COI 39 POY 
Strugnell et al. (2005) Coleoidea 16S, 12S, COI, rhod, pax-6, ODH 35 eye Bayesian, LogDet 
Takumiya et al. (2005) Coleoidea 12S, 16S, COI 36 SeqPup v. 0.9, ClustalX ver1.83 NJ, P, ML 
Akasaki et al. (2006) Coleoidea whole mitochondrial genome – ML 

*note these studies also used further information in some analyses in addition to gene sequences, e.g. morphology, allozymes, immunology etc.

Abbreviations: cyt b, cytochrome b apoenzyme, COI, cytochrome c oxidase subunit I; 16S, 16S rDNA; 12S, 12S rDNA; 28S, 28S rDNA; 18S, 18S rDNA; ODH, octopine dehydrogenase; rhod, rhodopsin; hist, histone H3; ef-1α, elongation factor-1α; All sequences were of partial fragments unless otherwise stated.

NJ, neighbour-joining; P, parsimony; ML, maximum likelihood.

Furthermore, since these first studies of cephalopod molecular phylogenetics, the range of sequence alignment and analysis methods available to phylogeneticists has increased (Table 1), and debate concerning the best methods to use has flourished (e.g. Wheeler, 1995, Kjer, Gillespie & Ober, 2007). Studies investigating cephalopod phylogenetics have aligned sequences by eye (Carlini & Graves, 1999; Carlini et al., 2000; Strugnell et al., 2004, 2005) or with the aid of alignment packages (Piertney et al., 2003; Yokobori et al., 2004; Zheng et al., 2004; Guzik et al., 2005) and have employed a variety of methods of analysis, including neighbour-joining (Allcock & Piertney, 2002; Warnke et al., 2003; Yokobori et al., 2004; Zheng et al., 2004), parsimony (Carlini & Graves, 1999; Anderson, 2000a,b; Carlini et al., 2000; Carlini, Young & Vecchione, 2001; Allcock & Piertney, 2002; Warnke et al., 2003; Lindgren et al., 2004, 2005; Nishiguchi et al., 2004; Zheng et al., 2004; Guzik et al., 2005), maximum likelihood (ML) (Anderson, 2000a,b; Carlini et al., 2000, 2001; Allcock & Piertney, 2002; Warnke et al., 2003; Strugnell et al., 2004; Yokobori et al., 2004; Guzik et al., 2005), Bayesian (Strugnell et al., 2004, 2005; Guzik et al., 2005) and LogDet (Anderson, 2000b; Strugnell et al., 2005). Recently, some studies have employed direct optimization where alignment is coupled with tree estimation in a dynamic procedure (Nishiguchi et al., 2004; Lindgren et al., 2004, 2005) (Table 1).

Although providing some insights [e.g. sister taxon relationships between the suborder Oegospida and family Bathyteuthidae (Strugnell et al., 2005)] none of these studies have conclusively resolved all higher-level cephalopod phylogenetic relationships and in many cases the results have been conflicting (see Akasaki et al., 2006; Nishiguchi & Mapes, 2007 for review of conflicting decapodiform relationships).

A number of reasons have been suggested for these varying and unresolved relationships. These include the early divergence of taxa, saturated sequence data, insufficient data, insufficient taxa and gene duplication (see Bonnaud et al., 1994, 1996; Carlini & Graves, 1999; Carlini et al., 2000; Lindgren et al., 2004; Strugnell et al., 2005 for discussion).

The large molecular data sets generated by Lindgren et al. (2004) (four genes) and Strugnell et al. (2004, 2005) (six genes) contained 18 of the same species (including 6 Octopodiformes and 11 Decapodiformes). Together, these provide the single largest dataset (with regard to sequence length) available for investigating higher-level phylogenetic relationships within the subclass Coleoidea. In the present study we used two methods to align these data: by eye and implied alignment using POY; and also three methods of analysis: parsimony, ML and Bayesian, to investigate the effect of these analyses on the resulting phylogeny. The effect of base composition heterogeneity upon coleoid phylogenetic relationships was also investigated.

MATERIAL AND METHODS

Eighteen species were used in the present study, including representatives from each higher-level taxon within the subclass Coleoidea (Table 2). Portions of nine genes were included, three mitochondrial genes (12S rDNA, 16S, rDNA, COI) and six nuclear genes (28S rDNA, 18S rDNA, histone, octopine dehydrogenase [ODH], pax-6 and rhodopsin). Sample details and methodologies used to obtain DNA sequences from these species are outlined in Lindgren et al. (2004) and Strugnell et al. (2005). Accession numbers for these sequences are listed in Table 2.

Table 2.

Accession numbers of each of the genes used in this study.

 Mitochondrial genes Nuclear genes 
 12S rDNA 16S rDNA COI 28S rDNA 18S rDNA hist. ODH pax-6 rhod. 
Nautiloidea          
 Nautilida          
  Nautilidae          
Nautilus pompilius AY616965 AY377628 AY557514 AF311688 AY557452   AY617039  
Coleoidea          
 Octopodiformes          
  Vampyromorpha          
   Vamyroteuthidae          
Vampyroteuthis infernalis AY545077 AY545101 AF000071 AY557548 AY557459 AY557408 AY545114 AY545139 AY545163 
  Octopoda          
   Allopsidae          
Haliphron atlanticus AY616942 AY616971 AY557516 AY557549 AY557460 AY557409 AY616910 AY617016 AY617040 
   Argonautidae          
Argonauta nodosa AY545080 AY545104 AY557517 AY557551 AY557462 AY557411 AY545117 AY545142 AY545166 
   Bolitaenidae          
Japetella diaphana AY545093 A252766 AY545192 AY557552 AY557463  AY545130 AY545155 AY545179 
   Octopodidae          
Eledone cirrhosa AY616946 AY616973 AY557520 AY557556 AY557467  AY616992 AY617020 AY617043 
Graneledone verrucosa AY545091 AY545111 AF000042 AY557557 AY557468 AY557413 AY545129 AY545153 AY545177 
 Decapodiformes          
  Sepiolida          
   Sepiolidae          
Heteroteuthis hawaiiensis AY616873 AY616884 AF000044 AY293703 AY557472 AY557416 AY616906 AY616937 AY616922 
  Sepiida          
   Sepiidae          
Sepia officinalis AY545098 X9570 AJ583491 AY557560 AY557471 AY557415 AY545135 AY545160 AF000947 
  Idiosepiida          
   Idiosepiidae          
Idiosepius pygmaeus AY545095 AJ001647 AY545193 AY293684 AY557477 AY557421 AY545132 AY5157 AY545181 
  Spirulida          
   Spirulidae          
Spirula spirula AY545097 AY293659 AY293709 AY557563 AY557476 AY557420 AY545134 AY545159 AY545183 
  Teuthida Myopsdia          
   Loliginidae          
Sepioteuthis lessoniana AY616869 AJ001649 AY131036 AY557566 AY557480 AY557424 AY616902 AY616933 AY616918 
  Teuthida Myopsdia          
   Bathyteuthidae          
Bathyteuthis abyssicola AY616958 AJ000104 AF000030 AY557568 AY557483 AY557427 AY617002 AY617032 AY617057 
Octopoteuthidae          
   Octopoteuthis nielseni    Cranchiidae AY616957 AY616983 AF000055 AY557591 AY557507  AY617011– AY617013 AY617031 AY617056 
Cranchia scabra    Ommastrephidae AY616962 DQ280046 AF000035 AY557571 AY557487 AY557430 AY617014 AY617015 AY617036 AY617061 
Illex coindetii AY616963 AY616985 AY617065 AY557593 AY557509 AY557450 AY617008 AY617015 AY617037 AY617062 
Sthenoteuthis oualaniensis AY545100 X79582 AF000069 AY557595 AY557511 AY557452 AY545137 AY545162 AY545185 
Ommastrephes batramii AY616866 AY616880 AF000057 AY557594 AY557510 AY557451 AY616899 AY616930 AY616915 
 Mitochondrial genes Nuclear genes 
 12S rDNA 16S rDNA COI 28S rDNA 18S rDNA hist. ODH pax-6 rhod. 
Nautiloidea          
 Nautilida          
  Nautilidae          
Nautilus pompilius AY616965 AY377628 AY557514 AF311688 AY557452   AY617039  
Coleoidea          
 Octopodiformes          
  Vampyromorpha          
   Vamyroteuthidae          
Vampyroteuthis infernalis AY545077 AY545101 AF000071 AY557548 AY557459 AY557408 AY545114 AY545139 AY545163 
  Octopoda          
   Allopsidae          
Haliphron atlanticus AY616942 AY616971 AY557516 AY557549 AY557460 AY557409 AY616910 AY617016 AY617040 
   Argonautidae          
Argonauta nodosa AY545080 AY545104 AY557517 AY557551 AY557462 AY557411 AY545117 AY545142 AY545166 
   Bolitaenidae          
Japetella diaphana AY545093 A252766 AY545192 AY557552 AY557463  AY545130 AY545155 AY545179 
   Octopodidae          
Eledone cirrhosa AY616946 AY616973 AY557520 AY557556 AY557467  AY616992 AY617020 AY617043 
Graneledone verrucosa AY545091 AY545111 AF000042 AY557557 AY557468 AY557413 AY545129 AY545153 AY545177 
 Decapodiformes          
  Sepiolida          
   Sepiolidae          
Heteroteuthis hawaiiensis AY616873 AY616884 AF000044 AY293703 AY557472 AY557416 AY616906 AY616937 AY616922 
  Sepiida          
   Sepiidae          
Sepia officinalis AY545098 X9570 AJ583491 AY557560 AY557471 AY557415 AY545135 AY545160 AF000947 
  Idiosepiida          
   Idiosepiidae          
Idiosepius pygmaeus AY545095 AJ001647 AY545193 AY293684 AY557477 AY557421 AY545132 AY5157 AY545181 
  Spirulida          
   Spirulidae          
Spirula spirula AY545097 AY293659 AY293709 AY557563 AY557476 AY557420 AY545134 AY545159 AY545183 
  Teuthida Myopsdia          
   Loliginidae          
Sepioteuthis lessoniana AY616869 AJ001649 AY131036 AY557566 AY557480 AY557424 AY616902 AY616933 AY616918 
  Teuthida Myopsdia          
   Bathyteuthidae          
Bathyteuthis abyssicola AY616958 AJ000104 AF000030 AY557568 AY557483 AY557427 AY617002 AY617032 AY617057 
Octopoteuthidae          
   Octopoteuthis nielseni    Cranchiidae AY616957 AY616983 AF000055 AY557591 AY557507  AY617011– AY617013 AY617031 AY617056 
Cranchia scabra    Ommastrephidae AY616962 DQ280046 AF000035 AY557571 AY557487 AY557430 AY617014 AY617015 AY617036 AY617061 
Illex coindetii AY616963 AY616985 AY617065 AY557593 AY557509 AY557450 AY617008 AY617015 AY617037 AY617062 
Sthenoteuthis oualaniensis AY545100 X79582 AF000069 AY557595 AY557511 AY557452 AY545137 AY545162 AY545185 
Ommastrephes batramii AY616866 AY616880 AF000057 AY557594 AY557510 AY557451 AY616899 AY616930 AY616915 

Sequence alignment and homology assessment

Two methods of sequence alignment were used within this study (1) by eye, and (2) using implied alignment using the homology scheme via POY (Wheeler, 2003; Giribet, 2005).

Aligned by eye

DNA sequences were compiled and aligned by eye in Se-Al v2.0a11 Carbon (Rambaut, 2002). Gaps were inserted where necessary to allow sequences to be aligned. Sequence data that were not alignable using this method were removed prior to analyses. Sequence alignment files are available on request. The total concatenated sequence length was 5,651 bp, of which 2,219 bp were variable.

Dynamic homology and implied alignments

Sequence data were analysed by using the direct optimization method described by Wheeler (1996) and implemented in the computer program POY. This method directly assesses the number of DNA sequence transformations (evolutionary events) required by a phylogenetic topology without the use of multiple sequence alignment. This is accomplished by generalization of existing character optimization procedures, including insertion and deletion events (indels) in addition to base substitutions. This method treats indels as processes, as opposed to the patterns implied by multiple sequence alignment (Wheeler, 1995). It is claimed that this method generates more efficient (and therefore simpler) explanations of sequence variation than multiple sequence alignment (Wheeler, 1996). Direct optimization, although computationally intense, is much less demanding than parsimony-based multiple sequence alignments when congruence among partitions is used as a criterion (Wheeler & Hayashi, 1998). The implied alignments produced via POY were used for both ML and Bayesian analyses. These sequences were concatenated for ML and Bayesian analysis (6,377 bp, of which 2,330 bp were variable).

Base composition heterogeneity

PAUP*4.0b10 (Swofford, 1998) was used for χ2 tests of composition homogeneity of the sequence data aligned by eye. Tests of base homogeneity were based on variable sites only. Where base composition heterogeneity was detected it was RY coded to remove base composition heterogeneity.

The three sequence data sets, (1) implied alignments, (2) aligned by eye, (3) aligned by eye and RY coded were analysed using three methods, (a) parsimony, (b) maximum likelihood, (c) Bayesian analysis. It is important to note that the sequence data aligned by eye were analysed using parsimony analyses in PAUP rather than POY.

Dynamic homology under parsimony

Molecular data were analysed with the computer program POY (Wheeler et al., 1996–2003) using the direct optimization method (Wheeler, 1996), with parsimony as the optimality criterion. Nodal support was calculated in POY using Farris's parsimony jackknifing procedure (Farris et al., 1996) for 100 replicates (using the commands: jackboot; replicates 100). Tree searches were conducted in parallel at Harvard University on a 19 dual-processor cluster (Darwin.oeb.harvard.edu) using pvm (parallel virtual machine). Commands for lad balancing of spawned jobs were used to optimize parallelization procedures (-parallel–dpm–jobspernode 2). Trees were built via a random-addition sequence procedure (10 replicates) followed by a combination of branch-swapping steps [SPR (subtree pruning and regrafting) and TBR (tree bisection and reconnection)] and tree fusing (Goloboff, 1999) in order to further improve on tree length minimization. Discrepancies between heuristic and actual tree length calculations were addressed by adjusting slop values (-slop5–checkslop10). Phylogenetic trees were obtained using parsimony with a gap/ts/tv cost of various weighting. Several analyses were implemented with character transformations weighted differently to determine how various phylogenetic hypotheses were affected (sensitivity analysis sensuWheeler, 1995). Each gene was analysed separately, using character transformations (indels/ts/tv) of equal weighting (111), and unequal weighting (121, 141, 211, 221, 241, 411, 421, 441). The parameter set that optimized the least amount of character incongruence was the equal weighted transformation (111) for all genes. Histone H3 and pax-6 were the two exceptions that also had similar character incongruence values for the 211 and 411 transformations. The final tree was drawn with Tree View (Win32) and consensus trees were analysed in PAUP version 4.02b (Swofford, 1998). To determine nodal support all jackknife calculations were performed in POY using the procedure described in Nishiguchi et al. (2004).

Implied alignment under parsimony

PAUP*4.0b10 (Swofford, 1998) was used to perform maximum parsimony analyses on the sequence data that were aligned by eye. All parsimony searches were performed with 1,000 random sequence-addition replicated and TBR (tree bisection-reconnection) branch swapping. All characters were unordered and equally weighted. One thousand bootstrap replicates were performed to measure the support for each clade on the phylogenetic trees.

Alignment by eye and implied alignment under maximum likelihood

PAUP*4.0b10 (Swofford, 1998) was used to perform 100 full heuristic searches. Starting trees were generated by the neighbour-joining method (NJ) (Saitou & Nei, 1987). A GTR + I + Γ likelihood model incorporating rate heterogeneity was used. Branch swapping was performed using TBR (tree-bisection-reconnection). Parameters were then re-estimated, and final branch swapping was performed using NNI (nearest-neighbour-interchange). ML bootstrap values of clade support were generated using the parameters estimated in the analysis, but with starting trees generated by the neighbour-joining method.

Alignment by eye and implied alignment under Bayesian analyses

MrBayes 3.1.2 (Ronquist & Huelsenbeck, 2003) was used to calculate marginal posterior probabilities using the GTR + I + Γ model of nucleotide substitution. Model parameter values were treated as unknown and were estimated in each analysis. Random starting trees were used for the analyses and were run between 1 and 500,000 generations, sampling the Markov chain every 100 generations.

Three strategies were used to ensure that analyses were not trapped in local optima: (1) analysis was performed twice, starting with a different random tree and log-likelihood values at stationarity were compared for convergence (Huelsenbeck & Bolback, 2001); (2) the topologies and clade posterior probabilities from each of the two analyses were compared for congruence (Huelsenbeck & Imennov, 2002); and (3) Metropolis-coupled Markov chain Monte Carlo (MCMCMC) was used with one cold and three incrementally heated Markov chains run simultaneously (default Mr Bayes heating values) to allow a more extensive exploration of parameter space (Huelsenbeck & Ronquist, 2001).

Stationarity was deemed to be reached when the average standard deviation of split frequencies, shown in MrBayes 3.1.2 was less than 0.01 (Ronquist & Huelsenbeck, 2003).

Tracer v1.3 (Rambaut & Drummond, 2003) was used to determine the correct ‘burnin-in’ for the analysis (i.e. the number of initial generations that must be discarded before stationarity is reached).

RESULTS

Sequence alignment

Alignment of the ODH, pax-6, COI and histone sequences required no insertion/deletion events (indels). Indels were introduced into aligned sequences of 12S rDNA, 16S rDNA, 28S rDNA, 18S rDNA and rhodopsin both by eye and ‘dynamically’ during the analysis using POY. The alignments of these genes where indels were required differed notably depending upon the alignment method (Table 3) (alignments available on request). A greater number of gaps were inserted using POY than by eye for the 12S rDNA, 16S rDNA and 28S rDNA genes (Table 3), whereas a greater number of insertions was used aligning by eye than by using POY for rhodopsin and 18S rDNA (Table 3). For each of these five genes requiring indels, regions that were deemed to be unalignable with confidence by eye were removed prior to analysis. In contrast, no sequence was removed from the POY analysis due to the fact that sequences are aligned simultaneously during analysis.

Table 3.

Comparison of alignment length of genes.

 Gene Total base pairs in gene sequenced (no gaps) (bp) Alignment method 
   POY (bp) Eye (total alignment length) (bp) Eye (unalignables removed, in analysis) (bp) 
Mitochondrial 12S rDNA 417 573 (417)* 486 (417)* 283 
 16S rDNA 528 627 (528)* 554 (528)* 427 
Nuclear 18S rDNA 2,845 1,893 (1,842)* 3,202 (2,845)* 1,943 
 28S rDNA 661 198 (191)* 166 (166)* 166 
 rhodopsin 1,040 1,022 (991)* 1,032 (954)* 765 
 Gene Total base pairs in gene sequenced (no gaps) (bp) Alignment method 
   POY (bp) Eye (total alignment length) (bp) Eye (unalignables removed, in analysis) (bp) 
Mitochondrial 12S rDNA 417 573 (417)* 486 (417)* 283 
 16S rDNA 528 627 (528)* 554 (528)* 427 
Nuclear 18S rDNA 2,845 1,893 (1,842)* 3,202 (2,845)* 1,943 
 28S rDNA 661 198 (191)* 166 (166)* 166 
 rhodopsin 1,040 1,022 (991)* 1,032 (954)* 765 

*Number in brackets indicates the starting sequence length without gaps. The portion of available sequence able to be aligned by eye was less for 28S and rhodopsin than by POY. A larger sequence fragment of 18S was attempted for alignment by eye, however a large proportion was unalignable and was removed prior to analysis.

Base composition heterogeneity

Chi-squared homogeneity tests of each of the genes shows that third positions of ODH, rhodopsin and COI have significant base frequency heterogeneity (Table 45). RY-coding the third positions of these genes was used to resolve base composition heterogeneity (Table 4). RY coding pools purines (adenine and guanine:R) and pyrimidines (cytosine and thymine:Y) into two-state categories (R,Y), and helps resolve bias resulting from differences in the relative frequency of either the two purines or pyrimidines (Phillips et al., 2001).

Table 4.

Chi-squared homogeneity test for base composition across all genes and codon positions.

Gene Codon position χ2(P
12S rDNA – 0.998 
16S rDNA – 0.997 
18S rDNA – 0.963 
28S rDNA – 1.000 
COI 1st 1.000 
COI 2nd 1.000 
COI 3rd 0.000 
COI (RY) 3rd 0.938 
histone H3 1st 1.000 
histone H3 2nd 1.000 
histone H3 3rd 0.560 
ODH 1st 1.000 
ODH 2nd 1.000 
ODH 3rd 0.000 
ODH (RY) 3rd 0.999 
pax-1st 1.000 
pax-2nd 1.000 
pax-3rd 0.945 
rhodopsin 1st 0.994 
rhodopsin 2nd 0.962 
rhodopsin 3rd 0.003 
rhodopsin (RY) 3rd 0.721 
Gene Codon position χ2(P
12S rDNA – 0.998 
16S rDNA – 0.997 
18S rDNA – 0.963 
28S rDNA – 1.000 
COI 1st 1.000 
COI 2nd 1.000 
COI 3rd 0.000 
COI (RY) 3rd 0.938 
histone H3 1st 1.000 
histone H3 2nd 1.000 
histone H3 3rd 0.560 
ODH 1st 1.000 
ODH 2nd 1.000 
ODH 3rd 0.000 
ODH (RY) 3rd 0.999 
pax-1st 1.000 
pax-2nd 1.000 
pax-3rd 0.945 
rhodopsin 1st 0.994 
rhodopsin 2nd 0.962 
rhodopsin 3rd 0.003 
rhodopsin (RY) 3rd 0.721 

Tests were performed on variable sites only. χ2(P) < 0.05 are in bold.

Table 5.

Phylogenetic relationships recovered by two alignment methods (by eye, dynamic homology/implied alignment using POY) and three analysis methods (P, parsimony; ML, maximum likelihood; Bayes, Bayesian).

Alignment method POY By eye 
    No RY RY 
Analysis method POY ML Bayes POY ML Bayes POY ML Bayes 
Vampyromorpha(Decapodiformes) √ √ √ √ √ √ √ 
Vampyromorpha(Octopoda) √ √ 
Octopoda √ √ √ √ √ √ √ √ √ 
Argonautoidea √ √ √ √ √ √ √ √ √ 
((Japetella, Graneledone)Eledone) √ √ √ √ √ 
(Eledone((Japetella,Graneledone)(Haliphron,Argonauta)) √ √ √ √ 
Decapodiformes √ √ √ √ √ √ √ √ √ 
(Oegopsida)(remaining Decapodiformes) √ √ √ √ 
Polyphyletic Sepioidea √ √ √ √ √ √ √ √ √ 
Ommastrephidae √ √ √ √ √ √ √ √ √ 
Oegopsida √ √ √ √ √ √ √ √ √ 
Bathyteuthoida(Oegopisda) √ √ √ √ √ √ √ √ √ 
Spirulida(Bathyteuthoida(Oegopisda)) √ √ √ √ 
Idiosepiidae(Sepioidea(Myopsida(Spirulida(Bathyteuthoida(Oegopisda))))) √ 
Sepioidea(Myopsida(Spirulida(Bathyteuthoida(Oegopisda)))) √ √ √ 
Myopsida(Spirulida(Bathyteuthoida(Oegopisda))) √ √ √ 
(Heteroteuthis, Idiosepius√ √ 
(Sepioteuthis, Idiosepius√ 
(Sepia, Idiosepius√ √ √ √ √ 
((Sepia, Idiosepius)Sepioeuthis√ √ √ √ 
(((Sepia, Idiosepius)Sepioeuthis)Spirula√ √ √ √ 
((((Sepia, Idiosepius)Sepioeuthis)Spirula)Heteroteuthis√ √ √ √ 
(Oegopsida,Bathyteuthoidea)(Sepioidea, Myopsida*) √ √ √ √ 
Alignment method POY By eye 
    No RY RY 
Analysis method POY ML Bayes POY ML Bayes POY ML Bayes 
Vampyromorpha(Decapodiformes) √ √ √ √ √ √ √ 
Vampyromorpha(Octopoda) √ √ 
Octopoda √ √ √ √ √ √ √ √ √ 
Argonautoidea √ √ √ √ √ √ √ √ √ 
((Japetella, Graneledone)Eledone) √ √ √ √ √ 
(Eledone((Japetella,Graneledone)(Haliphron,Argonauta)) √ √ √ √ 
Decapodiformes √ √ √ √ √ √ √ √ √ 
(Oegopsida)(remaining Decapodiformes) √ √ √ √ 
Polyphyletic Sepioidea √ √ √ √ √ √ √ √ √ 
Ommastrephidae √ √ √ √ √ √ √ √ √ 
Oegopsida √ √ √ √ √ √ √ √ √ 
Bathyteuthoida(Oegopisda) √ √ √ √ √ √ √ √ √ 
Spirulida(Bathyteuthoida(Oegopisda)) √ √ √ √ 
Idiosepiidae(Sepioidea(Myopsida(Spirulida(Bathyteuthoida(Oegopisda))))) √ 
Sepioidea(Myopsida(Spirulida(Bathyteuthoida(Oegopisda)))) √ √ √ 
Myopsida(Spirulida(Bathyteuthoida(Oegopisda))) √ √ √ 
(Heteroteuthis, Idiosepius√ √ 
(Sepioteuthis, Idiosepius√ 
(Sepia, Idiosepius√ √ √ √ √ 
((Sepia, Idiosepius)Sepioeuthis√ √ √ √ 
(((Sepia, Idiosepius)Sepioeuthis)Spirula√ √ √ √ 
((((Sepia, Idiosepius)Sepioeuthis)Spirula)Heteroteuthis√ √ √ √ 
(Oegopsida,Bathyteuthoidea)(Sepioidea, Myopsida*) √ √ √ √ 

*Myopsida falls within Sepioidea in this topology.

The data aligned by eye have been analysed for both nucleotide data and RY coded data. √, the relationship is supported, X, the relationship is not supported.

A number of taxonomic groupings are robust to the different methods of coding, alignment and analysis. The following taxa are always monophyletic: Octopoda, Argonautoidea, Ommastrephidae and Oegopsida (Figs 1–9). Furthermore, in each topology the Bathyteuthoida is the sister taxon to the Oegopsida (Figs 1–9). Bayesian posterior probabilities provide the highest support for each of these clades (Figs 3, 6, 9).

Figure 1.

Parsimony topology of coleoid cephalopod relationships obtained using direct optimization using POY. Jackknife support values are indicated beneath each node.

Figure 1.

Parsimony topology of coleoid cephalopod relationships obtained using direct optimization using POY. Jackknife support values are indicated beneath each node.

Figure 2.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were obtained from implied alignments using POY. Bootstrap support values are indicated beneath each node.

Figure 2.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were obtained from implied alignments using POY. Bootstrap support values are indicated beneath each node.

Figure 3.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were obtained from implied alignments using POY. Bayesian posterior probabilities are indicated beneath each node.

Figure 3.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were obtained from implied alignments using POY. Bayesian posterior probabilities are indicated beneath each node.

Figure 4.

Parsimony topology of coleoid cephalopod relationships. Sequences were aligned by eye. Boostrap support values are indicated beneath each node.

Figure 4.

Parsimony topology of coleoid cephalopod relationships. Sequences were aligned by eye. Boostrap support values are indicated beneath each node.

Figure 5.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye. Bootstrap support values are indicated beneath each node.

Figure 5.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye. Bootstrap support values are indicated beneath each node.

Figure 6.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye. Bayesian posterior probabilities are indicated beneath each node.

Figure 6.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye. Bayesian posterior probabilities are indicated beneath each node.

Figure 7.

Parsimony topology of coleoid cephalopod relationships. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Bootstrap support values are indicated beneath each node.

Figure 7.

Parsimony topology of coleoid cephalopod relationships. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Bootstrap support values are indicated beneath each node.

Figure 8.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Boostrap support values are indicated beneath each node.

Figure 8.

ML topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Boostrap support values are indicated beneath each node.

Figure 9.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Bayesian posterior probabilities are indicated beneath each node.

Figure 9.

Bayesian topology of coleoid cephalopod relationships obtained using GTR + I + Γ. Sequences were aligned by eye, and third positions of rhodopsin, COI and ODH were RY coded. Bayesian posterior probabilities are indicated beneath each node.

Vampyromorpha and the Decapodiformes are sister taxa in the phylogenies resulting from parsimony, ML and Bayesian analyses of the sequences aligned using POY and also by eye (no RY) with variable levels of support (Figs 1–6). This relationship is also recovered from parsimony analysis of RY coded sequence (BS = 99) (Fig. 7). In contrast, ML and Bayesian analysis of RY coded sequence aligned by eye recovered a sister-taxon relationship between Vampyromorpha and Octopoda, i.e. the Octopodiformes (Figs 8, 9). However, these relationships are not highly supported by bootstraps or posterior probabilities (Figs 8, 9).

The placement of Eledone within the Octopoda differs depending upon alignment and analysis method. Eledone is the sister taxon to a clade containing Japetella and Graneledone in each of the phylogenies resulting from the POY alignment, and also parsimony analysis of the sequence data aligned by eye, both RY coded and not RY coded (Figs 1–4,7). High support for this relationship is evident on each of these phylogenies. However, Eledone is highly supported as being basal within the Octopoda in analysis of sequence data aligned by eye (both RY coded and not RY coded) and analysed using ML (no RY, BS = 100; RY, BS = 100) and Bayesian analyses (no RY, PP = 1.00; RY, PP = 1.00) (Figs 5, 6, 8, 9).

Higher-level decapodiform relationships differ markedly between the various methods of alignment, coding, and analysis (Figs 1–9). Phylogenies generated from ML and Bayesian analyses of sequences aligned by eye (both RY coded and not RY coded) demonstrate Decapodiformes to be divided into two monophyletic groups, one containing the Oegopsida, and the second containing the remaining decapodiforms (i.e. Myopsida, Spirulidae, Sepiidae, Sepiolidae and Idiosepiidae) (Figs 5, 6,8,9). This division is highly supported by bootstrap support (no RY, BS = 98; RY, BS = 98) and posterior probabilities (no RY, PP = 0.99; RY, PP = 0.97) (Figs 5, 6,8,9). Within these topologies Sepia and Idiosepius are sister taxa, thereby rendering ‘Sepioidea’ (including Sepiidae, Sepiadariidae and Sepiolidae) polyphyletic (Figs 5, 6, 8, 9).

In contrast, a clade containing Heteroteuthis and Idiosepius is basal within decapodiforms in ML and Bayesian analysis (PP = 0.90) of sequence data aligned using POY (Figs 2, 3). Heteroteuthis alone is basal in phylogenies resulting from parsimony analysis of sequences aligned by eye, both RY coded (BS = 100) and not RY coded (BS = 100) (Figs 4, 7).

The position of Spirulidae within the Decapodiformes is highly dependent upon the method of alignment and analysis. Spirulidae are the sister taxon to a clade containing the Oegopsida and Bathyteuthoidea in all three analyses where sequences were aligned using POY, although support was only obtained for this relationship from the Bayesian analysis (PP = 1.00) (Figs 1–3). Interestingly, this same arrangement results from parsimony analysis of sequence data aligned by eye, not RY coded (BS = 84) (Fig. 4). In contrast, Spirulidae are the sister taxon to a clade containing Idiosepiidae, Sepiidae and Myopsida in the topologies resulting from ML and Bayesian analysis (PP = 0.99) of data aligned by eye (not RY coded) (Figs 5, 6) and in parsimony and Bayesian analysis (PP = 0.92) of RY coded data aligned by eye (Figs 8, 9).

DISCUSSION

The present study is the largest molecular analysis of cephalopod phylogeny to date, with regard to sequence length, and provides a thorough comparison of the effect of commonly used alignment and analysis methodologies on the resulting higher-level phylogenetic relationships.

The different alignment, analysis and coding methods used within this study produced a range of considerably different topologies. Only the clades Octopoda, Argonautoidea, Decapodiformes, Oegopsida, Ommastrephidae and a sister-taxon relationship between Bathyteuthidae and Oegopsida are robust to the alignment and analysis methods used.

Alignment methods

It is not surprising that different alignments can affect the resulting phylogeny, as the process of alignment aims to recover the evolutionary history of the sequences and therefore provides the very data upon which the algorithm performs (Giribet, Desalle & Wheller, 2002). For protein coding genes, the method of sequence alignment is usually insignificant, since in theory they should all produce the same alignment, i.e. an alignment without indels. However, as we demonstrate here, alignment methods for rDNAs, and both coding (i.e. rhodopsin) and non-coding genes can differ in their resulting sequence alignment and phylogenies.

There is debate in the literature regarding the best method of sequence alignment. Proponents of aligning sequences by eye (using secondary structural information) claim that they are ‘both philosophically and operationally superior’ (Kjer et al., 2007), whereas proponents of computational methods claim that alignments performed by eye are subjective and therefore not repeatable (Giribet & Wheeler, 1999).

Proponents of the method of direct optimization using POY claim that it avoids the problem of alignment by generalizing phylogenetic character analysis to include insertion/deletion events (indels), with the sequence data proceeding directly to phylogenetic reconstruction, obviating the necessity to create gap characters. Indels do not appear as states, but as transformations linking ancestral to descendent nucleotide sequences (Giribet & Wheeler, 1999; Giribet et al., 2002). POY assumes that shorter trees are better trees and that aligning nucleotides together based on state is parsimonious and algorithmically less costly. Kjer et al. (2007) argues that this is not justified in structurally conserved molecules such as rDNAs, where conserved structures in the molecules are more important than the states of the nucleotides.

There has been intense disagreement over the relative merits of manual alignment and direct optimization (Kjer, 1995; Wheeler, 1995; Shull et al., 2001; Belshaw & Quicke, 2002; Gillespie, Yoder & Wharton, 2005) but few rigorous comparisons of these methods. Recently, Kjer et al. (2007) compared the phylogenies obtained by three phylogeneticists who independently aligned and analysed the same 16S rDNA datset by eye (using rDNA secondary structure and analysed by parsimony) and using direct optimization within POY. Interestingly, although all three alignments by eye differed at some positions, each alignment produced nearly identical topologies. In contrast, when using POY, none of the three phylogeneticists converged on the same parameters or the same tree. Kjer et al. (2007) suggest that the reason for this is that gap cost to change ratios (used within POY) are arbitrary, and this allows different researchers to obtain different results.

Sequence alignments resulting from POY have been reported to be ‘gappy’ in some studies (Pons & Vogler, 2006) with the program inserting a greater number of indels than other methods when utilizing indel costs of 1. Similarly, in the present study POY inserted a greater number of gaps in the 12S rDNA, 16S rDNA and 28S rDNA alignments, even though several parameters were explored using POY (costs of 1, 2 and 4 for indels, transitions and transversions, respectively). In addition, no sequence data were removed from the POY alignments used in the analysis. In contrast, a notable proportion of the sequence alignments of 12S rDNA, 16S rDNA, 18S rDNA and rhodopsin was removed after alignment by eye as it was deemed to be unalignable and would contribute noisy signal to the analysis. Therefore the starting information present in both datasets differed. The sequence data that were deemed ‘unalignable’ when aligning by eye are by their nature ‘variable’ and would therefore have an important contribution in the POY alignments in determining the resulting phylogenetic relationships. Differences in phylogenetic relationships observed in this study between the two alignment methods is largely due to the deleted sequences. Over 60% of the sequence information in both the datasets was constant (i.e. not variable) further demonstrating the significance of these variable sites. Despite these obvious differences in output, both of these methods of sequence alignment are widely accepted and appear in the cephalopod (Table 1) and wider literature today. It is likely that debate will continue regarding the best method of sequence alignment and while this continues to be the case, it may be beneficial to employ more than one method of alignment in phylogenetic studies.

Analysis methods

There is considerable debate in the literature regarding methods of phylogenetic analysis (e.g. Giribet, 2003). Parsimony methods have the benefit of being relatively easy to understand and require few assumptions about the evolutionary process (Page & Holmes, 1998). However they have been shown to produce the wrong topology under the most realistic models of evolution (e.g. long branch attraction; Huelsenbeck & Hillis, 1993).

ML methods allow the incorporation of sophisticated models of sequence evolution and allow statistical tests of different evolutionary hypotheses (i.e. likelihood ratio testing Felsenstein, 1981) yet require very large computational resources. Furthermore ML methods have been shown to be susceptible to long branch repulsion and long branch attraction under some circumstances (Pol & Siddall, 2001).

Bayesian methodologies (differing from likelihood methods only in the use of a prior distribution of the quantity being inferred, which is typically the tree) have the advantage over ML methods of being computationally efficient. They allow very complex models of sequence evolution to be implemented and also can efficiently analyse large datasets. Bayesian methods have been criticized however, for producing unrealistically high posterior probability support (Suzuki, Glazko & Nei, 2002; Simmons, Pickett & Miya, 2004).

In the present study, the majority of topologies resulting from the three analysis methods on the implied aligned data (from POY) are very similar. The exception to this is the position of Idiosepius. In contrast, the method of analysis had a greater effect on the data aligned by eye. In many cases ML and Bayesian methods of analysis produced the same or very similar topology for both RY coded and non-RY coded data, while the parsimony analysis produced a different topology. This is the case for the relationships of octopod taxa, and the relationship between the Oegopsida and the rest of the decapodiforms. It is unsurprising that ML and Bayesian analysis methods produce more similar topologies than parsimony analysis, because both are based on the same probabilistic model of evolution. In contrast, parsimony analysis is based on the idea that the preferred phylogenetic tree is the one that requires the fewest evolutionary changes.

Discussion of phylogenetic relationships

Order Vampyromorpha

Vampyroteuthis infernalis is the only species within the order Vampyromorpha. It possesses a number of unusual characteristics including two pairs of fins in juveniles (one pair in adults) and a second pair of arms modified into retractile filaments. Traditionally Vampyromorpha and Octopoda have been suggested to be sister taxa due to embryological, developmental (Naef, 1928; Young & Vecchione, 1996; Boletzky, 2003) and morphological similarities, such as sperm morphology (Healy, 1989) and the presence of radial sucker symmetry (Lindgren et al., 2004). However, the vampyromorph gladius is known to be morphologically similar to that of decapodiforms (Toll, 1982, 1998). Previous molecular studies have found support for both a sister taxon relationship between Vampyromorpha and Octopoda (Bonnaud et al., 1997; Carlini & Graves, 1999; Lindgren et al., 2004; Strugnell et al., 2004,2005) and Vampyromorpha and the Decapodiformes (Bonnaud et al., 1997; Lindgren et al., 2004). This present study found support for both of these relationships. The majority of alignment and analysis combinations support a sister-taxon relationship between Vampyromorpha and Decapodiformes. Only ML and Bayesian analysis of the ‘by eye’ alignment of RY coded data support a sister-taxon relationship between Vampyromorpha and Octopoda. RY coding rectified the base composition heterogeneity identified in the third positions of COI, rhodopsin and ODH and thus is possible that this contributed to the Vampyromorpha and Octopoda sister-taxon relationship. RY coding also would have aided in reducing the effect of saturation (Phillips & Penny, 2003). However, parsimony analysis of the same dataset recovered a vampyromorph and decapodiform sister-taxon relationship. These results suggest that this relationship is unstable. The lineage Vampyromorpha is supposed to be at least 162 Myr from fossil evidence (Fischer & Riou, 2002) and has been estimated from fossil and molecular data to be potentially 252 Myr (Strugnell et al., 2006). The ancient diversification of this lineage provides support for the supposition that the molecular data used within this study are likely to be saturated at this level (Strugnell et al., 2005). Furthermore, the numerous extinction events throughout the Coleoidea during this time may contribute to the obscuring of affinities of Vampyromorpha (Lindgren et al., 2004).

Order Octopoda

Eledone was traditionally placed within the subfamily Eledoninae because it possesses an ink sac, a single row of suckers and large eggs (Robson, 1929). The taxonomic value of these characters has been debated; the presence of an ink sac has been suggested to be a function of depth (Robson, 1931; Voss, 1988; Allcock & Piertney, 2002) and sucker arrangement has been suggested to be a plastic character (Naef, 1921–1923; Voight, 1993a; Allcock & Piertney, 2002). Allcock & Piertney, (2002) suggested that sub-familial level assignment within the Octopodidae is ‘a totally artificial classification with no evolutionary basis.’ Eledone has been included in relatively few molecular studies (Bonnaud et al., 1997; Lindgren et al., 2004; Warnke et al., 2004). The present study recovered two differing placements for Eledone. All parsimony analyses, and also ML and Bayesian analyses of the POY alignment, show a sister-taxon relationship between Eledone and a clade containing Japetella and Graneledone, thus grouping together all species with a single row of suckers. In contrast ML and Bayesian analyses of data aligned by eye show Eledone to be basal within the Octopoda. This relationship was also recovered by Strugnell (2004), using a subset of the genes used within the present study, but with additional octopod species. Eledone possesses a number of morphological features supporting a basal position within the Octopoda, including the absence of a ligula (Naef, 1921–1923). It must be noted that there are relatively few octopod taxa included within the present study. The inclusion of additional taxa such as Benthoctopus, Bathypolypus and members of the suborder Cirrata would likely improve stability and resolution of octopod relationships.

Suborder Oegopsida and the family Bathyteuthidae

The suborder Oegospida contains squids that possess a gladius and lack a cornea. Molecular studies by Bonnaud et al. (1994, 1997), Carlini & Graves (1999), Carlini et al. (2000) and Lindgren et al. (2004) have suggested that the suborder may be polyphyletic, the later three studies reporting Spirula to fall within the Oegopsida. In contrast, Strugnell et al. (2005) supported a monophyletic Oegopsida. The present study also strongly supports a monophyletic Oegospida, since all alignment and analysis combinations supported this grouping. It is possible that the datasets in the previous studies that suggested a polyphyletic Oegopsida have been too small, and thus contained insufficient information to recover this relationship. All alignment and analysis combinations also support a sister-taxon relationship between the Oegopsida and the family Bathyteuthidae. This supports previous molecular studies by Carlini et al. (2000) and Strugnell et al. (2005) and also agrees with Naef's (1921–1923) suggestion that the Bathyteuthidae possess ‘primitive characters for all Oegopsida’.

Suborder Myopsida and Sepioidea

Traditionally Spirulidae, Sepiidae, Idiosepiidae and Sepiadariiae/Sepiolidae have been grouped together in the suborder Sepioidea (Naef, 1921–1923), while the suborder Myopsida was grouped with the suborder Oegopsida in the order Teuthoidea on the basis of similar gladii and tentacular clubs (Naef, 1916, 1921–1923). However, the Myopsida has also been suggested to be derived from the ‘Sepioidea’ line based on a number of characteristics including possession of a cornea, suckers with circularis muscle, beak without angle point and a vena cava ventral to the intestine (d'Orbigny, 1845; Berthold & Engeser, 1987; Engeser, 1997; Haas, 1997, see Young et al., 1998, for a more detailed discussion). Molecular studies have suggested a close relationship between the Myopsida and some or all members of the Sepioidea (Carlini et al., 2000; Lindgren et al., 2004; Strugnell et al., 2005), although the precise relationship has varied depending upon the genes and analyses used. The present study also suggests a closer relationship between the Myopsida and the Sepioidea than the Myopsida and the Oegopsida, although the exact configuration of this is dependent upon the alignment method and analysis employed. In the phylogenies resulting from data aligned using POY, Myopsida was consistently the sister taxon to a clade containing Spirulida, Bathyteuthoidea and Oegopsida, with the remaining Sepioidea taxa falling outside this clade. However, in the phylogenies resulting from ML and Bayesian analyses of data aligned by eye (RY coded and not RY coded) the Myopsida fell within Sepioidea, together forming a sister taxon to a clade containing the Oegopsida and Bathyteuthidae.

These results clearly show that differing alignment and analysis strategies commonly used in coleoid cephalopod phylogenetics can produce notably different phylogenetic relationships. Researchers are far from agreeing on a single ‘best’ strategy of phylogenetic analysis, because the advantages and disadvantages of competing strategies are not yet clear. Until such a time, we advocate the use of a variety of different alignment and analysis strategies in phylogenetic analysis.

ACKNOWLEDGEMENTS

We thanks the following people for their generosity in donating tissue samples: S. von Boletzky, David Carlini, Martin Collins, Stephen Craig, Eileen Dillane, F.G. Hochberg, T. Kubodera, W.K. Macy, Mark Norman, S. Piertney, Richard Stride, Michael Vecchione, Kerstin Warnke and Richard Young. We are also grateful to G. Giribet for running the POY analysis on the Darwin cluster. J.S. is supported by a Natural Environment Research Council Antarctic Funding Initiative grant (NE/C506321/1). M.K.N. is partially supported by NIH SO6 GM008136-32S2 and NSF DEB-0316516.

REFERENCES

AKASAKI
T.
NIKAIDO
M.
TSUCHIYA
K.
SEGAWA
S.
HASEGAW
M.
OKADA
N.
Extensive mitochondrial gene arrangements in coleoid Cephalopoda and their phylogenetic implications
Molecular Phylogenetics and Evolution
 , 
2006
, vol. 
38
 (pg. 
648
-
658
)
ALLCOCK
A.L.
PIERTNEY
S.B.
Evolutionary relationships of Southern Ocean Octopodidae (Cephalopoda: Octopda) and a new diagnosis of Pareledone
Marine Biology
 , 
2002
, vol. 
140
 (pg. 
129
-
135
)
ANDERSON
F.E.
Phylogenetic relationships among loliginid squids (Cephalopoda: Myopsida) based on analyses of multiple data sets
Zoological Journal of the Linnean Society
 , 
2000
, vol. 
a 130
 (pg. 
603
-
633
)
ANDERSON
F.E.
Phylogeny and historical biogeography of the Loliginid squids (Mollusca: Cephalopoda) based on mitochondrial DNA sequence data
Molecular Phylogenetics and Evolution
 , 
2000
, vol. 
b 15
 (pg. 
191
-
214
)
BELSHAW
R.
QUICKIE
D.L.J.
Robustness of ancestral state estimates: Evolution of life strategy in ichneumonoid parasitoids
Systematic Biology
 , 
2002
, vol. 
51
 (pg. 
450
-
477
)
BERTHOLD
T.
ENGESER
T.
Phylogenetic analysis and systematization of the Cephalopoda (Mollusca)
Verhandlungen des naturwissenschaftlichen Vereins Hamburg
 , 
1987
, vol. 
29
 (pg. 
187
-
220
)
BOLETZKY
S.von
Evolutionary aspects of development, life style, and reproductive mode in incirrate octopods (Mollusca, Cephalopoda)
Revue suisse de Zoologie
 , 
1992
, vol. 
99
 (pg. 
755
-
770
)
BOLETZKY
S.von
Biology of early life stages in cephalopod molluscs
Advances in Marine Biology
 , 
2003
, vol. 
44
 (pg. 
144
-
184
)
BONNAUD
L.
BOUCHER-RODONI
R.
MONNEROT
M.
Phylogeny of decapod Cephalopods based on partial 16S rDNA nucleotide sequences
Comptes Rendus de l'Académie des Sciences
 , 
1994
, vol. 
317
 (pg. 
575
-
580
)
BONNAUD
L.
BOUCHER-RODONI
R.
MONNEROT
M.
Relationship of some coleoid cephalopods established by 3' end of the 16S rRNA and cytochrome oxidase III gene sequence comparison
American Malacological Bulletin
 , 
1996
, vol. 
12
 (pg. 
87
-
90
)
BONNAUD
L.
BOUCHER-RODONI
R.
MONNEROTT
M.
Phylogeny of cephalopods inferred from mitochondrial DNA sequences
Molecular Phylogenetics and Evolution
 , 
1997
, vol. 
7
 (pg. 
44
-
54
)
BONNAUD
L.
RODHOUSE
P.G.
BOUCHER-RODONI
R.
A phylogenetic study of the squid family Onychoteuthidae (Cephalopoda: Oegopsida)
Proceedings of the Royal Society of London. Series B
 , 
1998
, vol. 
265
 (pg. 
1761
-
1770
)
BOUCHER-RODONI
R.
BONNAUD
L.
Biochemical and molecular approach to cephalopod phylogeny
American Malacological Bulletin
 , 
1996
, vol. 
12
 (pg. 
79
-
85
)
CARLINI
D.B.
GRAVES
J.E.
Phylogenetic analysis of cytochrome c oxidase I sequences to determine higher-level relationships within the coleoid cephalopods
Bulletin of Marine Science
 , 
1999
, vol. 
64
 (pg. 
57
-
76
)
CARLINI
D.B.
REECE
K.S.
GRAVES
J.E.
Actin gene family evolution and the phylogeny of coleoid cephalopods (Mollusca: Cephalopoda)
Molecular Biology and Evolution
 , 
2000
, vol. 
17
 (pg. 
1353
-
1370
)
CARLINI
D.B.
YOUNG
R.E.
VECCHIONE
M.
A molecular phylogeny of the octopoda (Mollusca:Cephalopoda) evaluated in light of morphological evidence
Molecular Phylogenetics and Evolution
 , 
2001
, vol. 
21
 (pg. 
388
-
397
)
CLARKE
M.R.
Clarke
M.R.
Trueman
E.R.
Evolution of recent cephalopods-A brief review
The Mollusca. Paleontology and neontology of cephalopods
 , 
1988
, vol. 
Vol. 12
 
San Diego
Academic Press
(pg. 
1
-
355
)
D'ORBIGNY
A.
Mollusques vivants et fossiles
 , 
1845
, vol. 
2
 
Paris
ENGESER
T.
The fossil Coleoidea page
1997
 
FARRIS
J.S.
ALBERT
V.A.
KÄLLERSJÖ
M.
LIPSCOMB
D.
KLUGE
A.G.
Parsimony jacknifing outperforms neighbor-joining
Cladistics
 , 
1996
, vol. 
12
 (pg. 
99
-
124
)
FELSENSTEIN
J.
Evolutionary trees from DNA sequences: a maximum likelihood approach
Journal of Molecular Evolution
 , 
1981
, vol. 
17
 (pg. 
368
-
376
)
FISCHER
J.C.
RIOU
B.
Vampyronassa rhodanica nov. gen. nov sp., vampyromorphe (Cephalopoda, Coleoidea) du Callovien inférieur de la Voulte-sur-Rhône (Ardèche, France)
Annales de Paléontologie
 , 
2002
, vol. 
88
 (pg. 
1
-
17
)
GILLESPIE
J.J.
YODER
M.J.
WHARTON
R.A.
Predicted secondary structures for 28S and 18S rRNA from Ichneumonoidea (Insecta: Hymenoptera: Apocrita): impact on sequence alignment and phylogeny estimation
Journal of Molecular Evolution
 , 
2005
, vol. 
61
 (pg. 
114
-
137
)
GIRIBET
G.
Stability in phylogenetic formulations, and its relationship to nodal support
Systematic Biology
 , 
2003
, vol. 
52
 (pg. 
554
-
564
)
GIRIBET
G.
Generating implied alignments under direct optimization using POY
Cladistics
 , 
2005
, vol. 
21
 (pg. 
396
-
402
)
GIRIBET
G.
DESALLE
R.
WHEELER
W.C.
DeSalle
R.
Giribet
G.
Wheeler
W.C.
‘Pluralism’ and the aims of phylogenetic research
Molecular systematics and evolution: theory and practice
 , 
2002
Basel
Birkhauser
(pg. 
141
-
146
)
GIRIBET
G.
WHEELER
W.C.
On gaps
Molecular Phylogenetic and Evolution
 , 
1999
, vol. 
13
 (pg. 
132
-
143
)
GOLOBOFF
P.A.
Analyzing large data sets in reasonable times: solutions for composite optima
Cladistics
 , 
1999
, vol. 
15
 (pg. 
S26
-
S34
)
GRIMPE
G.
Teuthologische Mitteilungen. VII:Systematiche Ubersicht der Nordsee-Cephalopoden
Zoologischer Anzeiger
 , 
1921
, vol. 
52
 (pg. 
297
-
305
)
GUZIK
M.T.
NORMAN
M.D.
CROZIER
R.H.
Molecular phylogeny of the benthic shallow-water octopuses (Cephalopoda: Octopodinae)
Molecular Phylogenetic and Evolution
 , 
2005
, vol. 
37
 (pg. 
235
-
248
)
HAAS
W.
Der Ablauf der Entwicklungsgeschichte der Decabrachia (Cephalopoda, Coleoidea)
Palaeontographica
 , 
1997
, vol. 
245A
 (pg. 
63
-
81
)
HEALY
J.M.
Spermatozoa of the deep-sea cephalopod Vampyroteuthis infernalis Chun: ultrastructure and possible phylogenetic significance
Philosophical Transactions of the Royal Society, Series B
 , 
1989
, vol. 
323
 (pg. 
589
-
600
)
HUELSENBECK
J.P.
BOLBACK
J.P.
Empirical and hierchical Bayesian estimation of ancestral states
Systematic Biology
 , 
2001
, vol. 
50
 (pg. 
351
-
366
)
HUELSENBECK
J.P.
HILLIS
D.M.
Success of phylogenetic methods in the four-taxon case
Systematic Biology
 , 
1993
, vol. 
40
 (pg. 
247
-
264
)
HUELSENBECK
J.P.
IMENNOV
N.S.
Geographic origin of mitochondrial DNA: accomodating phylogenetic uncertainty and model comparison
Systematic Biology
 , 
2002
, vol. 
51
 (pg. 
155
-
165
)
HUELSENBECK
J.P.
RONQUIST
F.
MrBayes:Bayesian inference of phylogeny
Bioinformatics
 , 
2001
, vol. 
17
 (pg. 
754
-
755
)
KHROMOV
D.N.
Cuttlefishes in the systematics and phylogeny of Cephalopoda
Zoologicheskii Zhurnal
 , 
1990
, vol. 
69
 (pg. 
12
-
20
)
KJER
K.M.
Use of rRNA secondary structure in phylogenetic studies to identify homologous positions: an example of alignment and data presentation from the frogs
Molecular Phylogenetics and Evolution
 , 
1995
, vol. 
4
 (pg. 
314
-
330
)
KJER
K.M.
GILLESPIE
J.J.
OBER
K.A.
Opinions on multiple sequence alignment, and an empirical comparison of repeatability and accuracy between POY and structural alignment
Systematic Biology
 , 
2007
, vol. 
56
 (pg. 
133
-
146
)
LINDGREN
A.R.
GIRIBET
G.
NISHIGUCHI
M.K.
A combined approach to the phylogeny of Cephalopoda (Mollusca)
Cladistics
 , 
2004
, vol. 
20
 (pg. 
454
-
486
)
LINDGREN
A.R.
KATUGIN
O.N.
AMEZQUITA
E.
NISHIGUCHI
M.K.
Evolutionary relationships among squids of the family Gonatidae (Mollusca: Cephalopoda) inferred from three mitochondrial loci
Molecular Phylogenetics and Evolution
 , 
2005
, vol. 
26
 (pg. 
101
-
111
)
NAEF
A.
Uber neue Sepioliden aus dem Golf von Neapel
Pubblicazioni della Stazione Zooloica di Napoli
 , 
1916
, vol. 
1
 (pg. 
1
-
10
)
NAEF
A.
Cephalopoda
 , 
1921
 
–1923 Fauna e flora del Golfo di Napoli, Monograph. (translated from German by the Israel program for Scientific translations, Jerusalem 1972)
NAEF
A.
Die Cephalopoden. Vol. 2. Embryologie
 , 
1928
Monographie
Fauna e Flora del Golfo di Napoli
pg. 
35
 
NISHIGUCHI
M.K.
MAPES
R.
Ponder
W.
Lindberg
D.
Cephalopoda
Molluscan evolution
 , 
2007
Berkeley, California
University of California Press
(pg. 
161
-
197
)
NISHIGUCHI
M.K.
LOPEZ
J.E.
BOLETZKY
S.von
Enlightenment of old ideas from new investigations: more questions regarding the evolution of bacteriogenic light organs in squids
Evolution and Development
 , 
2004
, vol. 
6
 (pg. 
41
-
49
)
PAGE
R.D.M.
HOLMES
E.C.
Molecular evolution. A phylogenetic approach.
 , 
1998
Oxford
Blackwell Science
PHILLIPS
M.J.
LIN
Y.-H.
HARRISON
G.L.
PENNY
D.
Mitochondrial genomes of a bandicoot and a brushtail possum confirm the monophyly of australidelphian marsupials
Proceedings of the Royal Society of London, Series B
 , 
2001
, vol. 
268
 (pg. 
1533
-
1538
)
PHILLIPS
M.J.
PENNY
D.
The root of the mammalian tree inferred from whole mitochondrial genomes
Molecular Phylogenetics and Evolution
 , 
2003
, vol. 
28
 (pg. 
171
-
185
)
PICKFORD
G.E.
The Vampyromorpha. A new order of dibranchiate Cephalopoda
Věstník Československé zoologické společnosti
 , 
1939
, vol. 
6–7
 (pg. 
346
-
358
)
PIERTNEY
S.B.
HUDELOT
C.
HOCHBERG
F.G.
COLLINS
M.A.
Phylogenetic relationships among cirrate octopods (Mollusca: Cephalopoda) resolved using mitochondrial 16S ribosomal DNA sequences
Molecular Phylogenetics and Evolution
 , 
2003
, vol. 
27
 (pg. 
348
-
353
)
POL
D.
SIDDALL
M.E.
Biases in maximum likelihood and parsimony: a simulation approach to a 10-taxon case
Cladistics
 , 
2001
, vol. 
17
 (pg. 
266
-
281
)
PONS
J.
VOGLER
A.P.
Size, frequency, and phylogenetic signal of multiple-residue indels in sequence alignment of introns
Cladistics
 , 
2006
, vol. 
22
 (pg. 
144
-
156
)
RAMBAUT
A.
2002
Oxford University
 
Se-Al. Ver. v2.0a11 Carbon
RAMBAUT
A.
DRUMMOND
A.J.
2003
Oxford University
 
Tracer. v1.0.1
ROBSON
G.C.
A monograph of Recent Cephalopoda. Part I. Octopodinae
 , 
1929
London
British Museum (Natural History)
ROBSON
G.C.
A monograph of the Recent Cephalopoda. Part II. The Octopoda (excluding the Octopodinae)
 , 
1931
London
British Museum (Natural History)
RONQUIST
F.
HUELSENBECK
J.P.
MrBayes 3: Bayesian phylogenetic inference under mixed models
Bioinformatics
 , 
2003
, vol. 
19
 (pg. 
1572
-
1574
)
SAITOU
N.
NEI
M.
The neighbor-joining method. a new method for reconstructing phylogenetic trees
Molecular Biology and Evolution
 , 
1987
, vol. 
4
 (pg. 
406
-
425
)
SHULL
V.L.
VOGLER
A.P.
BAKER
M.D.
MADDISON
D.R.
HAMMOND
P.M.
Sequence alignment of 18S ribosomal RNA and the basal relationships of adephagan beetles: evidence for monophyly of aquatic families and the placement of Trachypachidae
Systematic Biology
 , 
2001
, vol. 
50
 (pg. 
945
-
969
)
SIMMONS
M.P.
PICKETT
K.M.
MIYA
M.
How meaningful are Bayesian support values?
Molecular Biology and Evolution
 , 
2004
, vol. 
21
 (pg. 
188
-
199
)
STRUGNELL
J.
The molecular evolutionary history of the Class Cephalopoda (Phylum Mollusca)
 , 
2004
Department of Zoology, University of Oxford
 
DPhil thesis
STRUGNELL
J.
JACKSON
J.
DRUMMOND
A.J.
COOPER
A.
Divergence time estimates for major cephalopod groups: evidence from multiple genes
Cladistics
 , 
2006
, vol. 
22
 (pg. 
89
-
96
)
STRUGNELL
J.
NORMAN
M.
DRUMMOND
A.J.
JACKSON
J.
COOPER
A.
Molecular phylogeny of coleoid cephalopods (Mollusca: Cephalopoda) using a multigene approach; the effect of data partitioning on resolving phylogenies in a Bayesian framework
Molecular Phylogenetics and Evolution
 , 
2005
, vol. 
37
 (pg. 
426
-
441
)
STRUGNELL
J.M.
NORMAN
M.D.
DRUMMOND
A.J.
COOPER
A.
The octopuses that never came back to earth: neotenous origins for pelagic octopuses
Current Biology
 , 
2004
, vol. 
18
 (pg. 
R300
-
R301
)
SUZUKI
Y.
GLAZKO
G.V.
NEI
M.
Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics
Proceedings of the National Academy of Science of the USA
 , 
2002
, vol. 
99
 (pg. 
16138
-
16143
)
SWOFFORD
D.L.
PAUP*4.0-Phylogenetic Analysis Using Parsimony (*and other methods)
 , 
1998
Sunderland, MA
Sinauer
TAKUMIYA
M.
KOBAYASHI
M.
TSUNEKI
K.
FURUYA
H.
Phylogenetic relationships among major species of Japanese coleoid cephalopods (Mollusca: Cephalopoda) using three mitochondrial DNA sequences
Zoological Science
 , 
2005
, vol. 
22
 (pg. 
147
-
155
)
TOLL
R.B.
The comparative morphology of the gladius in the order Teuthodea (Mollusca: Cephalopoda) in relation to systematics and phylogeny
 , 
1982
University of Miami
 
PhD thesis
TOLL
R.B.
The gladius in teuthoid systematics
International workshop on the systematics and biogeography of cephalopods
 , 
1998
Washington D.C
Smithsonian Institution Press
(pg. 
65
-
68
)
VOIGHT
J.R.
The arrangement of suckers on octopodid arms as a continuous character
Malacologia
 , 
1993
, vol. 
a 35
 (pg. 
351
-
359
)
VOIGHT
J.R.
A cladistic reassessment of Octopodid classification
Malacologia
 , 
1993
, vol. 
b 35
 (pg. 
343
-
349
)
VOIGHT
J.R.
Cladistic analysis of the octopods based on anatomical characters
Journal of Molluscan Studies
 , 
1997
, vol. 
63
 (pg. 
311
-
325
)
VOSS
G.L.
Nixon
M.
Messenger
J.B.
Present status and new trends in cephalopod systematics
The biology of cephalopods
 , 
1977
London
Academic Press
(pg. 
49
-
60
)
VOSS
G.L.
Clarke
M.R.
Trueman
E.R.
Evolution and phylogenetic relationships of deep-sea octopods (Cirrata and Incirrata)
The Mollusca
 , 
1988
, vol. 
12
 
San Diego
Academic Press
(pg. 
253
-
276
Paleontology and neontology of cephalopods
WARNKE
K.
PLÖTNER
J.
SANTANA
J.I.
RUEDA
M.J.
LLINÀS
O.
Reflections on the phylogenetic position of Spirula (Cephalopoda): preliminary evidence from the 18S ribosomal RNA gene
Berliner Paläobiologisch Abhandlungen
 , 
2003
, vol. 
3
 (pg. 
253
-
260
)
WARNKE
K.
SOLLER
R.
BLOHM
D.
SAINT-PAUL
U.
A new look at geographic and phylogenetic relationships within the species group surrounding Octopus vulgaris (Mollusca, Cephalopoda): indications of very wide distribution from mitochondrial DNA sequences
Journal of Zoological Systematics and Evolutionary Research
 , 
2004
, vol. 
42
 (pg. 
306
-
312
)
WHEELER
W.C.
Sequence alignment, parameter sensitivity, and the phylogenetic analysis of molecular data
Systematic Biology
 , 
1995
, vol. 
44
 (pg. 
321
-
331
)
WHEELER
W.C.
Optimization alignment: the end of multiple sequence alignment in phylogenetics?
Cladistics
 , 
1996
, vol. 
12
 (pg. 
1
-
9
)
WHEELER
W.C.
Implied alignment: a synapomorphy-based multiple-sequence alignment method and its use in cladogram search
Cladistics
 , 
2003
, vol. 
19
 (pg. 
261
-
268
)
WHEELER
W.C.
HAYASHI
C.Y.
The phylogeny of extant chelicerate orders
Cladistics
 , 
1998
, vol. 
14
 (pg. 
173
-
192
)
WHEELER
W.C.
GLADSTEIN
D.S.
LAET
J.D.
1996
 
–2003 POY, v.3.0.11 ftp.amnh.org/pub/molecular/poy Wheeler
YOKOBORI
S.-I.
FUKUDA
N.
NAKAMURA
M.
AOYAMA
T.
OSHIMA
T.
Long-term conservation of six duplicated structural genes in cephalopod mitochondrial genomes
Molecular Biology and Evolution
 , 
2004
, vol. 
22
 (pg. 
2034
-
2046
)
YOUNG
R.E.
VECCHIONE
M.
Analysis of morphology to determine primary sister-taxon relationships within celeoid cephalopods
American Malacological Bulletin
 , 
1996
, vol. 
12
 (pg. 
91
-
112
)
YOUNG
R.E.
VECCHIONE
M.
DONOVAN
D.T.
The evolution of coleoid cephalopods and their present biodiversity and ecology
South African Journal of Marine Science
 , 
1998
, vol. 
20
 (pg. 
393
-
420
)
ZHENG
X.
YANG
J.
LIN
X.
WANG
R.
Phylogenetic relationships among the decabrachia cephalopods inferred from mitochondrial DNA sequences
Journal of Shellfish Research
 , 
2004
, vol. 
23
 (pg. 
881
-
886
)