ARTICLE IN PRESS MOLECULAR PHYLOGENETICS AND EVOLUTION Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx www.elsevier.com/locate/ympev Molecular systematics of Salmonidae: combined nuclear data yields a robust phylogeny Bernard J. Crespi* and Michael J. Fulton Behavioural Ecology Research Group, Department of Biosciences, Simon Fraser University, 8888 University Drive, Burnaby, BC, Canada V5A 1S6 Received 12 May 2003; revised 15 August 2003 Abstract The phylogeny of salmonid ﬁshes has been the focus of intensive study for many years, but some of the most important relationships within this group remain unclear. We used 269 Genbank sequences of mitochondrial DNA (from 16 genes) and nuclear DNA (from nine genes) to infer phylogenies for 30 species of salmonids. We used maximum parsimony and maximum likelihood to analyze each gene separately, the mtDNA data combined, the nuclear data combined, and all of the data together. The phylogeny with the best overall resolution and support from bootstrapping and Bayesian analyses was inferred from the combined nuclear DNA data set, for which the diﬀerent genes reinforced and complemented one another to a considerable degree. Addition of the mitochondrial DNA degraded the phylogenetic signal, apparently as a result of saturation, hybridization, selection, or some combination of these processes. By the nuclear-DNA phylogeny: (1) (Hucho hucho, Brachymystax lenok) form the sister group to (Salmo, Salvelinus, Oncorhynchus, H. perryi); (2) Salmo is the sister-group to (Oncorhynchus, Salvelinus); (3) Salvelinus is the sistergroup to Oncorhynchus; and (4) Oncorhynchus masou forms a monophyletic group with O. mykiss and O. clarki, with these three taxa constituting the sister-group to the ﬁve other Oncorhynchus species. Species-level relationships within Oncorhynchus and Salvelinus were well supported by bootstrap levels and Bayesian analyses. These ﬁndings have important implications for understanding the evolution of behavior, ecology and life-history in Salmonidae. Ó 2003 Published by Elsevier Inc. Keywords: Salmonidae; Phylogeny; Total evidence; Anadromy 1. Introduction The family Salmonidae comprises three subfamilies, Coregoninae (whiteﬁsh and ciscoes), Thymallinae (grayling), and Salmoninae (char, trout, and salmon). The most speciose of these, Salmoninae, includes ﬁve genera distributed throughout the Northern Hemisphere, Brachymystax (lenok), Hucho (huchen and taimen), Oncorhynchus (Paciﬁc trout and salmon), Salmo (Atlantic salmon and brown trout), and Salvelinus (char) (Hart, 1973; Hendry and Stearns, 2003; Scott and Crossman, 1973). Salmonid ﬁshes have long been of great interest due to the commercial and recreational value of some species, and they are becoming increasingly important as model systems for addressing a wide range of evolutionary and ecological questions (Elliot, * Corresponding author. Fax: +604-291-3496. E-mail address: [email protected] (B.J. Crespi). 1055-7903/$ - see front matter Ó 2003 Published by Elsevier Inc. doi:10.1016/j.ympev.2003.08.012 1994; Groot and Margolis, 1991; Hendry and Stearns, 2003). Inference of a robust phylogeny for this group is important for comparative analyses of salmonid adaptations (e.g., Crespi and Teo, 2002; Fleming, 1998), comparative genomics (e.g., Woram et al., 2003), studies involving inference of ancestral states (e.g., McDowall, 1997; McLennan, 1994; Stearley, 1992), and evaluation of conservation priorities (Crandall et al., 2000). Despite the importance of salmonids to humans, and to terrestrial and marine ecosystems, their evolutionary history has remained a matter of considerable dispute for many years (e.g., Domanico et al., 1997; McKay et al., 1996; McPhail, 1997; Norden, 1961; Oakley and Phillips, 1999; Phillips and Oakley, 1997; Phillips and Pleyte, 1991; Regan, 1914; Utter et al., 1973; Utter and Allendorf, 1994). Previous species-level and genus-level phylogenetic research on salmonid ﬁshes have provided insights into some relationships, but numerous questions remain, most notably the among-genus diver- ARTICLE IN PRESS 2 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx gences, and species-level relationships within Oncorhynchus and Salvelinus. The lack of a comprehensive, well-resolved and wellsupported phylogeny for Salmonidae can be largely attributed to previous studies using relatively small subsets of extant salmonid diversity, and only one or at most several genes or other character sets (e.g., morphology or karyology). To overcome these limitations, we have assembled and analyzed all available DNA-sequence data for the species in this family. The main goals of our study are twofold: (1) to use these data to infer the best tree for the family as a whole, and for particular lineages; and (2) to assess what additional data (i.e., sequence from which genes) are needed to achieve a species-level tree for the entire group. 2. Methods 2.1. Data set We compiled all of the available sequence data for salmonid ﬁshes and one outgroup (Plecoglossus altivelis) (Salmoniformes: Osmeridae), which comprised 269 sequences of mitochondrial DNA (from 16 genes) and nuclear DNA (from 9 genes) for 31 species (Table 1). The bulk of these data were from Genbank, with several additional sequences graciously provided to us by T. Oakley and R. Phillips. Some species for which very little data were available (e.g., only one or several genes) were not included. However, some species with substantial amounts of missing data were included, as inclusion of such taxa has been shown to increase phylogenetic accuracy and is not expected to produce misleading results (Wiens and Reeder, 1995; Wiens, 1998a). Complete mitochondrial DNAs were available for C. lavaretus, O. mykiss, O. tshawytscha, P. altivelis, S. salar, Sv. alpinus, and Sv. fontinalis. The MHC genes used were chosen randomly, one for each species, from the larger sample of alleles in Genbank. The sequences were aligned gene by gene using Clustal X (Thompson et al., 1997) and by eye, and regions with ambiguous alignments (e.g., parts of the DLOOP) were excluded. The full data set had 27,593 base pairs, and it is available as a NEXUS ﬁle from BC. 2.2. Phylogenetic analyses We used maximum parsimony and maximum likelihood in PAUP (Swoﬀord, 2002) and Bayesian analysis in MrBayes (http://brahms.biology.rochester.edu/ software.html) (Alfaro et al., 2003; Hall, 2001; Huelsenbeck et al., 2001; Rannala and Yang, 1996; Yang and Rannala, 1997) for our analyses. We analyzed each gene separately (for genes with at least 13 taxa repre- sented, and for vitellogenin and MHC), the full mitochondrial data set, the full nuclear-gene data set, and all of the data combined. For most of the analyses of mitochondrial data, P. altivelis was used as the outgroup. However, because data from this species were not available for nuclear genes, C. lavaretus, B. lenok, or H. perryi were also used as outgroups, depending on which was available and closest to the ingroup based on previous studies. MrMODELTEST (http:// www.ebc.uu.se/systzoo/staﬀ/nylander.html) was used to choose the most appropriate models of molecular evolution for the likelihood analyses of each separate gene, and for the combined data sets (Posada and Crandall, 1998, 2001; Sullivan and Swoﬀord, 2001). To assess the robustness of the inferred trees, we used bootstrapping with 500 replicates under maximum parsimony and bootstrapping with 200 replicates for maximum likelihood. Maximum likelihood bootstrapping was not computationally feasible for the combined data sets. For the Bayesian analyses, we used at least 5000 trees after stabilization of the likelihoods to compute the a posteriori probabilities, which can be interpreted as the probabilities that particular clades are correct. These probabilities tend to be less conservative than maximum-likelihood bootstrap values (Alfaro et al., 2003; Douady et al., 2003). Although they tend to identify more correct monophyletic groups than do parsimony or likelihood bootstrapping in simulations, Bayesian support values may also overestimate the degree of clade support, especially for lineages descending from short internodes (Alfaro et al., 2003; Douady et al., 2003). We have taken a Ôconditional combinationÕ approach (Bull et al., 1993; De Queiroz et al., 1995; Huelsenbeck et al., 1996) to analyze data derived from multiple genes and loci. This approach involves an assessment of congruence, using various means, prior to a decision to combine data sets or analyze them separately. We follow this approach for two reasons. First, some genes or loci may produce inconguent and incorrect results, due to such processes as sampling error, hybridization, natural selection, rate variation among lineages, variation in base composition, or a high degree of saturation (Sanderson and Schaﬀer, 2002; Slowinksi and Page, 1999). Second, given only minor eﬀects from such processes in a combined data set, a total evidence analysis should yield the best results, because diﬀerent genes should provide resolution and support in diﬀerent regions of the tree (Bull et al., 1993; Huelsenbeck et al., 1996). Our analysis of congruence was constrained by the variable sets of taxa for which data were available for each gene, which precluded direct comparisons of genes on a pairwise basis with ML tests (e.g., Huelsenbeck and Bull, 1996) or other methods (see Barker and Lutzoni, 2002). As a result, we evaluated the degree of congru- ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 1 Genbank sequences used in the analyses B. lenok C. artedi C. autumnalis C. clupeaformis C. kiyi C. lavaretus H. hucho H. perryi O. clarki O. gorbuscha O. keta O. kisutch O. masou O. mykiss O. nerka O. rhodurus O. tshawytscha P. altivelis P. coulteri P. williamsoni S. orhidana S. salar S. trutta Sv. alpinus Sv. conﬂuentus Sv. fontinalis Sv. leucomaenis Sv. malma Sv. namaycush T. arcticus T. thymallus DLOOP 12S 16S AF125519 AF246932 AJ250996 AF239253 U95191 AB034824 AF125513 AF125513 ND2 CO1 AB034824 AB034824 L29771 L29771 L29771 AF246933 AB034824 AF254863 AB034824 AB034824 AF296347 AF296345 AF296344 AF296342 AF125510 L29771 AF296343 AF254865 AB039901 AF318037 AF429780 L29771 U59926 AF113119 AF392054 AB047553 AY008713 AY008696 AF392054 AB047553 AF392054 AB047553 AF392054 AB047553 AF392054 AB047553 AF392054 AB047553 AF133701 U62286 AF154851 AF133701 AF133701 AF133701 AF154851 AF126004 AF154850 AF154851 AF133701 AF117718 AF154851 AF154851 AF133701 M64917 AF154851 AF154850 AF060445 AF154850 AF154850 AF154850 AF076906 AF036381 AF076908 AF036381 Atp6 CO3 ND3 ND4L AF154850 AF297988 AF298043 AF297989 AF329990 AF329989 CO2 B. lenok C. artedi C. autumnalis C. clupeaformis C. kiyi C. lavaretus H. hucho H. perryi O. clarki O. gorbuscha O. keta O. kisutch O. masou O. mykiss O. nerka O. rhodurus O. tshawytscha P. altivelis P. coulteri P. williamsoni S. orhidana S. salar S. trutta Sv. alpinus Sv. conﬂuentus Sv. fontinalis Sv. leucomaenis Sv. malma Sv. namaycush ND1 AF113117 L29771 Atp8 AF246934 AJ133367 AB034824 AB034824 AB034824 D84147 L29771 L29771 D63336 L29771 AF392054 AB047553 AF392054 AB047553 AF392054 AB047553 AB034824 AB034824 AF294830 AF294831 D84147 AF294829 D63336 L29771 AF294832 AF312575 AF055090 AF055089 AF055092 U28364 L29771 AF055091 U28363 AF392054 AB047553 AF392054 AB047553 AB034824 D84147 D63336 L29771 AF392054 AB047553 AJ133369 AF133701 AF133701 AF133701 AF154851 AF133701 X76247 AF154851 AF154851 AF154850 AF133701 AF154851 AF133701 U61181 AF154851 AF154850 AF154850 AF154850 AF154850 AF154850 U61182 AF154851 3 ARTICLE IN PRESS 4 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 1 (continued) CO2 Atp8 Atp6 CO3 ND3 ND4L ND4 ND5 ND6 Cyt b GH1c GH2c AF125052 AF125213 AF005919 AF005917 AF005920 AF005924 AF005926 AF005927 AF005925 AB001865 AF005907 AF005908 AF005913 AF075572 L04688 U04931 AF005923 U14551 J03797 U14535 Oakley AF005914 T. arcticus T. thymallus B. lenok C. artedi C. autumnalis C. clupeaformis C. kiyi C. lavaretus H. hucho H. perryi O. clarki O. gorbuscha O. keta O. kisutch O. masou O. mykiss O. nerka O. rhodurus O. tshawytscha P. altivelis P. coulteri P. williamsoni S. orhidana S. salar S. trutta Sv. alpinus Sv. conﬂuentus Sv. fontinalis Sv. leucomaenis Sv. malma Sv. namaycush T. arcticus T. thymallus AJ251592 AB034824 AB034824 AY032633 U66039 AY032633 U66039 AF125051 L29771 L29771 AF125050 L29771 AF392054 AB047553 AF392054 AB047553 AF392054 AB047553 AF133701 AF133701 AF133701 AF154851 AF154851 AF154851 AF154850 AF154850 AF154850 ITS1 AB034824 AF172397 D58396 AY032633 AF165077 AF165078 AF165079 D58403 L29771 AJ314568 AF392054 AB047553 AY008700 AY008701 AF202033 AF133701 X77526 AF154851 ITS2 Oakley AF005921 AF005915 M21573 AF005912 AF005909 AF005911 AF154850 D58397 AF319544 AF270689 AF270858 GH2d B. lenok C. artedi C. autumnalis C. clupeaformis C. kiyi C. lavaretus H. hucho H. perryi O. clarki O. gorbuscha O. keta O. kisutch O. masou O. mykiss O. nerka O. rhodurus O. tshawytscha P. altivelis P. coulteri P. williamsoni S. orhidana S. salar S. trutta Sv. alpinus Sv. conﬂuentus Sv. fontinalis AB034824 18s AF005922 Oakley AF005910 GH1d VIT MHC AF243426 AB001865 L04688 J03797 U14535 AF454745 M94900 AF174612 AF170535 Oakley Oakley AF170536 AF170533 Oakley AF170539 AF170538 AF170540 AF170542 AF170543 AF170537 AF170534 AF170541 AF030250 AF243427 AF243428 AF454747 AJ011689 U14551 AF454748 U34717 U34703 U34692 U34697 U34715 U34711 U34719 AY008709 AY008695 M21573 AF201313 AF201312 AF072862 AF059899 M94902 M94903 AF174609 AF174613 AF174611 AJ427629 X98839 AF469620 Phillips Phillips Phillips Phillips Phillips X70166 AF454750 AF454751 AF454752 ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx 5 Table 1 (continued) CO2 Sv. leucomaenis Sv. malma Sv. namaycush T. arcticus T. thymallus Atp8 Atp6 M94904 M94905 M94906 AF174607 AF174608 AF174610 CO3 ND3 ND4L Phillips Phillips Phillips AF454753 Fig. 1. Bootstrap majority-rule phylogenies inferred from individual-gene DNA data sets. MP ¼ maximum parsimony, ML ¼ maximum likelihood, BAY ¼ Bayesian analysis. GH ¼ growth hormone, ITS ¼ internal transcribed spacer, VIT ¼ vitellogenin, MHC ¼ major histocompatibility complex. ML bootstraps were not computationally feasible for the DLOOP data set. ARTICLE IN PRESS 6 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Fig. 1. (continued) ence between the results from analyses of separate genes via inspection of bootstrap or a posteriori probability values, to determine how many diﬀerent genes and loci supported a given monophyletic group and to what extent, and to identify any strongly supported nodes that diﬀered among genes (De Queiroz et al., 1995). Given potential incongruence in one or more parts of a phylogeny, we agree with Wiens (1998b) that combining data may still represent the best strategy for inferring the most-accurate tree, subject to the caveat that the gene or genes involved in possible incongruence should be considered questionable and may require removal from a combined analysis. In addition to analyzing each gene separately, we also compared the degrees of resolution and support obtained from analyses of the combined mitochondrial data set, the combined nuclear data set, the combined nuclear data set without the MHC data (since salmonid MHC is known to be under strong balancing selection, Miller and Withler, 1996), all of the data except MHC, and all of ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx 7 Fig. 1. continued) the data combined. Cunningham (1997) has shown that congruence and phylogenetic accuracy tend to be positively correlated, such that strongly supported combined-data trees are likely to reﬂect congruence among trees from the individual data sets. Finally, alternative a priori hypotheses for the placement of particular species and sets of species were also evaluated using the SH test (Shimodaira and Hasegawa, 1999) and the Templeton test (Templeton, 1983), implemented in PAUP* as described below. 3. Results 3.1. Phylogenetic analyses of individual and combined data Bootstrap majority-rule consensus trees from analyses using maximum parsimony and maximum likelihood, and a posteriori clade support values from Bayesian analysis, are shown in Fig. 1 for each of the individual genes, and Fig. 2 shows maximum-parsimony ARTICLE IN PRESS 8 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Fig. 1. (continued) bootstraps and Bayesian support for the combined data sets. The individual-gene trees diﬀer considerably in the species included and the degree of support for various relationships. We assessed the degree of support for the main phylogenetic hypothesis among Salmonidae by collating the bootstrap and Bayesian a posteriori values relevant to each putative monophyletic group of interest (Table 2). This analysis allows us to assess the degree to which the results from the diﬀerent data sets are con- gruent, reinforce one another, or provide conﬂicting signal. 3.2. Oncorhynchus The monophyly of Oncorhynchus, which is strongly supported by virtually all previous studies (Murata et al., 1993, 1996, 1998; Oakley and Phillips, 1999; Oleinik, 1997; Phillips and Oakley, 1997 and references therein), ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx 9 Fig. 2. Bootstrap majority-rule and Bayesian phylogenies for the combined data sets: (A) all mtDNA, (B) all nuclear DNA excluding MHC, (C) all nuclear DNA, (D) all nuclear DNA and mtDNA excluding MHC, (E) all data. was also supported in our analyses by all of the individual genes and by the combined data sets (Table 2). However, the strength of support varied considerably across data sets, with the nuclear genes tending to provide higher bootstrap and Bayesian a posteriori values than the mitochondrial genes, and the combined mtDNA data set (Fig. 2A) returning the weakest support. The monophyly of Oncorhynchus excluding O. clarki, O. masou, and O. mykiss was strongly supported here by GH1C and by the combined nuclear DNA. By contrast, this group was not supported by analyses of ITS1, CO3, ND3, and GH2C, the other genes that provided any degree of bootstrap majority-rule resolution for this clade. However, the monophyly of this group was not strongly contradicted by these analyses, as the relevant bootstrap values were low. The reduced support for this clade, compared to the genus as a whole, is due primarily to the presence of O. mykiss, O. clarki or O. masou among the other Oncorhynchus in some of the trees. The clades (O. nerka, O. gorbuscha, O. keta) and (O. gorbuscha, O. keta) have been inferred in most previous studies of Oncorhynchus (Domanico and Phillips, 1995; Kitano et al., 1997; McKay et al., 1996; Murata et al., 1993, 1996, 1998; Oakley and Phillips, ARTICLE IN PRESS 10 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 2 Bootstrap values from maximum likelihood (200 replicates, in plain text) or Bayesian analyses (majority-rule of at least 5000 trees, in italics), and maximum parsimony (500 replicates, in parentheses) analyses, for relevant clades in Figs. 1 and 2 All Oncorhynchus (Paciﬁc salmon and Paciﬁc trout) (O. tshawytscha, O. kisutch, O. nerka, O. gorbuscha, O. keta, O. mykiss, O. clarki, O. masou) GH1C 100 100 (100) ITS1 100 100 (100) ITS2 100 100 (100) GH2C 96 100 (93) CYTB 70 93 (98) CO3 71 84 (96) ND3 86 52 (81) 16S 82 97 (59) MHC (62) All mt DNA 76 (61) All nuclear DNA (excl. MHC) 100 (100) All nuclear DNA 100 (99) All nuclear DNA (excl. MHC) + mtDNA 100 (100) All nuclear + mt DNA 100 (100) Paciﬁc salmon excluding O. masou (O. tshawytscha, O. kisutch, O. nerka, O. gorbuscha, O. keta) GH1C 98 100 (98) ITS2 55 GH2C Not monophyletic CO3 Not monophyletic ITS1 Not monophyletic CYTB Not monophyletic ND3 Not monophyletic All nuclear DNA (excl. MHC) 100 (94) All nuclear DNA 100 (92) All nuclear DNA (excl. MHC) + mtDNA Not monophyletic 54 93 (66) 60 99 73 95 (56) 79 100 (includes O. masou, 99) (O. nerka, O. gorbuscha, O. keta) GH1C GH2C MHC ITS2 CYTB ITS1 ND3 CO3 16S All mt DNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 100 100 (100) 98 100 (93) 69 99 (74) 62 86 (75) 72 65 84 includes O. kisutch 87 100 (92) includes O. kisutch Not monophyletic 60 99 (monophyletic, 56) Not monophyletic O. nerka with O. tshawytscha, O. kisutch 55,63 77,88 (55,71) 92 (85) 100 (100) 100 (100) 100 (100) 100 (100) (O. gorbuscha, O. keta) CYTB ITS2 GH1C GH2C MHC ITS1 ND3 CO3 All mt DNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 94 98 (99) 69 91 58 95 (69) 54 53 52 86 50 (51) (Not monophyletic O. gorbuscha with O. nerka, O. kisutch: 57) Not monophyletic 60 99 (monophyletic, 66) 97 (94) 100 (92) 100 (90) 100 (100) 100 (100) (O. tshawytscha, O. kisutch) MHC VIT 100 100 (99) 99 100 (99) ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 2 (continued) ITS2 CYTB GH1C CO3 16S ITS1 GH2C ND3 All mt DNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 90 100 (88) 91 99 (95) 62 85 (73) Not monophyletic 60 99 Not monophyletic, includes O. nerka: 55,63 77,88 (55,71) Not monophyletic 73 95 (56) Not monophyletic 78 100 (76) Not monophyletic 87 100 (92) Not monophyletic 88 (75) 100 (72) 100 (99) Not monophyletic 100 (90) 100 (59) Paciﬁc trout and O. masou (O. mykiss, O. masou) ITS2 CO3 ITS1 ND3 MHC DLOOP CYTB All mt DNA All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mtDNA 76 80 (90) 78 54 76 (56, includes O. tshawytscha) Not monophyletic 52 Not monophyletic 81 95 (62) Not monophyletic 83 Not monophyletic 98 Not monophyletic 98 (90) (O. mykiss with O. clarki) Not monophyletic 100 (O. mykiss with O. clarki) Not monophyletic 100 (100) (O. mykiss with O. clarki) Not monophyletic 100 (100) (O. mykiss with O. clarki) (O. mykiss, O. clarki) GH1C CYTB GH2C DLOOP ND3 CO3 All mt DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear DNA All nuclear + mt DNA 100 100 (100) 84 98 (95) 84 100 (87) 83 Not monophyletic 50 50 Not monophyletic 60 99 98 (90) 100 (100) 100 100 (100) (O. mykiss, O. clarki, O. masou) ND3 CO3 CYTB ND3 All All All All nuclear DNA (excl. MHC) nuclear DNA nuclear DNA (excl. MHC) + mtDNA nuclear + mt DNA O. masou as sister-taxon to all other Oncorhynchus; Not monophyletic 52 Not monophyletic 60 99 ML, MP: O. masou basal but unresolved in Oncorhynchus ML, MP: O. masou basal but unresolved in Oncorhynchus Not monophyletic 79 Not monophyletic 50; O. masou basal in Oncorhynchus; (MP: O. masou sister to other Oncorhynchus, 71) 100 (95) 100 (79) Not monophyletic 99 (71), O. masou with O. tshawytscha 96, Not monophyletic (51), O. masou basal in Paciﬁc salmon Salvelinus (Sv. alpinus, Sv. malma, Sv. conﬂuentus, Sv. leucomaenis, Sv. fontinalis, Sv. namaycush) GH1C 100 100 (100) ITS1 100 100 (100) ITS2 100 100 (100) GH2C 92 100 (94) GH2D 57 69 (55) All nuclear DNA (excl. MHC) 100 (100) All nuclear DNA 100 (100) All nuclear DNA (excl. MHC) + mtDNA 100 (100) All nuclear + mt DNA 100 (100) 11 ARTICLE IN PRESS 12 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 2 (continued) (Sv. alpinus, Sv. malma, Sv. conﬂuentus, Sv. leucomaenis) ITS2 ITS1 GH1C GH2D GH2C All mtDNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 91 99 (97) 59 84 (54) Not monophyletic Not monophyletic Not monophyletic Not monophyletic 99 (75) 100 (73) 99 Not monophyletic (Sv. alpinus, Sv. malma) ITS1 GH2C GH2D DLOOP GH1C ITS2 All mt DNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 96 100 (99) 87 100 (94) 88 92 (79) (63) 52 Not monophyletic 96 100 (96) 87 (52) 100 (100) 100 (100) 100 (100) 100 (100) (Sv. conﬂuentus, Sv. leucomaenis) ITS1 GH1C GH2C ITS2 GH2D All mt DNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 98 100 (96) 58 97 (71) 60 82 (55) Not monophyletic 96 100 (97) Not monophyletic 68 79 (62) Not monophyletic 87 (75) 100 (72) 100 (73) 100 (74) 100 (73) (Sv. fontinalis, Sv. namaycush) ITS2 GH1C ITS1 GH2D GH2C ND3 All mtDNA All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 100 100 (100) 66 86 (52) (52) Not monophyletic Not monophyletic Not monophyletic Not monophyletic 100 (98) 100 (99) 100 (75) 100 (77) Sister-group to Oncorhynchus (Oncorhynchus, Salvelinus) GH1C VIT ND3 DLOOP CO3 CYTB All nuclear DNA (excl. MHC) All nuclear DNA All nuclear DNA (excl. MHC) + mtDNA All nuclear + mt DNA 100 100 (100) 81 100 (92) 85 100 (81) 78 (includes P. coulteri, B. lenok, Oncorhynchus paraphyletic) 51 54 (67) 58 (53) (Salvelinus lineage includes Hucho and B. lenok) 100 (94) 100 (93) 64, Salvelinus includes Hucho and B. lenok (61) 91 (57) (Oncorhynchus, Salmo) ITS1 96 100 (96) 50 68 76 87 (monophyletic: 52) 79 (62) 100 (82) (75) 51 (50) 68 79 (62) 76 100 (82) 83 (89) 87 Results from genes for which relationships remain unresolved in the 50% majority-rule consensus trees are not shown. For clades that are not monophyletic, the highest bootstrap or a posteriori Bayesian value supporting the lack of monophyly is listed. Genes are listed from top to bottom in order of the degree to which they tend to support (top) or contradict (bottom) the presence of the monophyletic group shown. ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx 1999; Oohara et al., 1997, 1999; Osinov, 1999; Phillips and Oakley, 1997; Shedlock et al., 1992 and references therein) and have yet to be strongly contradicted. In our analyses, we found support for (O. nerka, O. gorbuscha, O. keta) from the nuclear genes GH1C, GH2C, MHC, and ITS2, as well as from each of the combined data sets. The results that were incongruent with (O. nerka, O. gorbuscha, O. keta) tended to involve relatively low bootstrap and Bayesian values, excepting those from ND3 and GH2D. The group (O. gorbuscha, O. keta) was accorded bootstrap and Bayesian support from CYTB, ITS2, GH1C, GH2C, ITS1, and MHC though most of these support values were rather low. However, the contradictory results from ND3 and CO3 were weak, and the combined data sets provided very strong support for this clade. The sister-taxon status of O. tshawytscha with O. kisutch is supported by considerable previous phylogenetic work, and has yet to be contradicted strongly (Du et al., 1993; Kitano et al., 1997; McKay et al., 1996; Murata et al., 1993, 1996, 1998; Oakley and Phillips, 1999; Oohara et al., 1997, 1999; Osinov, 1999; Phillips and Oakley, 1997 and references therein). This clade was supported in our analyses by MHC, VIT, ITS2, CYTB, GH1C, the combined nuclear data sets, and the complete data set. By contrast, analyses of CO3, 16S, ITS1, GH2C, ND3, the mtDNA data, and the complete data set without MHC, yielded results contradictory to the presence of this clade. However, the bootstrap and Bayesian values for these genes tended to be relatively low compared to the clade support values from the genes that supported it. Moreover, the phylogenetic positions of these two species were inconsistent across CO3, 16S, ITS1, GH2C, and ND3 (Fig. 1), such that there was no evidence for one or more speciﬁc alternative placements. Considered together, these results are consistent with the sister-taxon status of O. tshawytscha 13 and O. kisutch, although support for this grouping is not entirely unambiguous. The phylogenetic placement of O. masou, in relation to the other Oncorhynchus, has not received strong support from previous analyses (e.g., Kitano et al., 1997; McKay et al., 1996; Murata et al., 1996; Oleinik, 2000; Oohara et al., 1997, 1999). To elucidate the relationships of O. mykiss, O. clarki and O. masou among themselves and to the other Oncorhynchus, we jointly evaluated the degrees of support for the clades (O. mykiss, O. clarki), (O. mykiss, O. masou), and (O. mykiss, O. clarki, O. masou) (Table 2). The sister-taxon relationship of O. mykiss with O. clarki was clearly upheld by the analyses of GH1C, CYTB, GH2C, the mtDNA data, and the full combined data set, and it was relatively weakly incongruent with results from ND3 and CO3. In addition, a relationship of O. mykiss with O. masou was inferred from ITS2 and (marginally) from ITS1. Considering all three species together, we note that by most singlegene analyses, O. masou appears relatively basal in Oncorhynchus but its position is unresolved, as are the positions of O. mykiss, O. clarki, or both in the analyses of CO3, 16S, and ND3. A basal but weakly resolved position for O. masou is also apparent in analyses of the combined mtDNA data (Fig. 2A), the full data set with MHC excluded (Fig. 2D), and the the full data set (Fig. 2E) using maximum parsimony. The clearest results for the position of O. masou come from analyses of the nuclear DNA data set, with or without MHC (Fig. 2B and C). These analyses provided strong bootstrap support, Bayesian support, or both for the monophyly of (O. mykiss, O. clarki, O. masou). The Bayesian analyses also strongly supported sister-taxon status of O. mykiss and O. clarki. These results indicate that the various individual genes provide no support or weak to moderate support for the clade (O. masou, (O. mykiss, O. clarki)), and that no evidence strongly Table 3 Results from Templeton and SH tests that in each case compare the best (unconstrained) tree to an alternative, constraint tree: (a) best unconstrained tree vs. best tree showing O. masou not with O. mykiss and O. clarki, (b) best unconstrained tree vs. best tree with Salmo as sister-group to Oncorhynchus, (c) best unconstrained tree vs. best tree where O. kisutch and O. tshawytscha are not sister-taxa Constraint Data set Nuclear data without MHC MP ML All nuclear data MP ML (a) Not (O. masou, O. mykiss, O. clarki) (b) (Salmo, Oncorhynchus) (c) Not (O. kisutch, O. tshawytscha) 1703 vs. 1709 P ¼ 0:058 22712.21 vs. 22729.14 P ¼ 0:030 1703 vs. 1718 P ¼ 0:022 22712.21 vs. 22733.51 P ¼ 0:046 1700 vs. 1702 P > 0:40 22712.21 vs. 22718.89 P > 0:10 1911 vs. 1914 P > 0:40 25687.05 vs. 25696.96 P ¼ 0:087 1911 vs. 1926 P ¼ 0:022 25687.05 vs. 25708.15 P ¼ 0:050 1908 vs. 1919 P ¼ 0:012 25687.05 vs. 25717.39 P ¼ 0:018 For Templeton tests (MP), tree lengths are shown, and for SH tests (ML), likelihoods are shown. For the MP analysis of (c), H. hucho was excluded due to apparent long-branch attraction into Oncorhynchus in the constraint tree. Qualitatively similar results were obtained using the constraint tree ‘‘Not (Oncorhynchus, Salvelinus)’’. ARTICLE IN PRESS 14 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx contradicts these relationships. However, taken together, the nuclear ones reinforce and complement one another suﬃciently to provide strong evidence that O. masou belongs with the Paciﬁc trout. 3.3. Salvelinus The monophyly of Salvelinus was supported by all of the genes for which data on this genus was available, though support from GH2D, and from the combined mitochondrial DNA data set, was relatively weak (Table 2). Some of the relationships within Salvelinus were reasonably clear, but for others diﬀerent genes yielded incongruent results. Thus, the monophyly of (Sv. alpinus, Sv. malma, Sv. conﬂuentus, Sv. leucomaenis) was supported strongly by ITS2, moderately supported by the complete nuclear DNA data set, and weakly supported by ITS1 and GH1C. However, the diﬀering results from analyses of GH2D and GH2C, the combined mtDNA data set, and the full data set, preclude unambiguous interpretation of this result. Similarly, the groups (Sv. alpinus, Sv. malma), (Sv. conﬂuentus, Sv. leucomaenis), and (Sv. fontinalis, Sv. namaycush) were each upheld, often with high bootstrap and Bayesian support values, for one or more genes, but analyses of one or more other genes yielded notably incongruent results. For each of these cases, the combined data sets, especially the nuclear data, lent strong or moderate support to the group, but this support tended to stem from only one or two genes that may or may not be indicating the correct phylogeny. The simplest interpretation of these heterogeneous results is that one or more of the genes does not reﬂect the species tree for Salvelinus. In particular, the data from ITS2 disrupts the monophyly of both (Sv. alpinus, Sv. malma) and (Sv. conﬂuentus, Sv. leucomaenis), because it strongly groups Sv. conﬂuentus with Sv. alpinus and Sv. malma, and the data from GH2C, GH2D, and ND3 prevents (Sv. alpinus, Sv. malma, Sv. conﬂuentus, Sv. leucomaenis) and (Sv. fontinalis, Sv. namaycush) from each being monophyletic, because they position Sv. namaycush strongly with Sv. alpinus and Sv. malma. As discussed below, these ﬁndings are consistent with the hypothesis that extensive hybridization has obfuscated relationships among species of Salvelinus as inferred from DNA-sequence data. 3.4. Relationship of Oncorhynchus to Salvelinus and Salmo A sister-taxon relationship between Oncorhynchus and Salvelinus is strongly supported by GH1C, VIT, ND3, and the combined nuclear data sets, and weakly supported by CO3 and CYTB (Table 2). By contrast, the ITS1 data groups Oncorhynchus and Salmo with such a high degree of conﬁdence that is it clearly incongruent with the results from these other data sets. This analysis of ITS1 appears problematic, however, because our analyses of other genes have shown that H. perryi may not be an appropriate outgroup for an analysis of intergeneric relationships between Oncorhynchus, Salmo, and Salvelinus. CYTB groups H. perryi with Salvelinus with 57% bootstrap conﬁdence in the parsimony analysis, and GH1C and the complete nuclear DNA data set group H. perryi with Salmo, with 90–92% and 66% conﬁdence respectively. Thus, given that H. perryi may belong in the ingroup, the analysis of ITS1 cannot be used to address relationships between Oncorhynchus, Salmo, and Salvelinus. The ﬁsh species that are closest to this set of taxa, but deﬁnitely not in the ingroup, are Cichlidae, which are highly divergent in ITS1 (i.e., on the order of 50% or more divergent in nucleotide sequence). Use of Neochromis nigricans (Genbank U67338) as an outgroup, aligned to the taxa in the ITS1 data set using Clustal X, yielded a phylogeny with Oncorhynchus and Salmo as sister-taxa but with only 58% bootstrap support from maximum parsimony analysis (500 replicates). 3.5. Statistical tests of alternative hypotheses We used SH tests and Templeton tests to evaluate alternative hypotheses for three important questions in salmonid phylogenetics: (1) the phylogenetic position of O. masou, (2) the sister-group to Oncorhynchus, and (3) the monophyly of (O. tshawytscha, O. kisutch) (Table 3). These tests used the two combined nuclear-gene data sets, which, as described below, provide what we believe is the best estimate of salmonid phylogeny. The monophyly of (O. masou, O. mykiss, O. clarki) was statistically supported by the SH test of the nuclear data without MHC, and support was marginally nonsigniﬁcant (0:05 < P < 0:10) for the Templeton test of this data set and the SH test of the full nuclear data set (Table 3). These results are consistent with erosion of bootstrap and Bayesian support for (O. masou, O. mykiss, O. clarki) with the addition of the MHC data (Fig. 2). A sister-taxon relationship between Oncorhynchus and Salmo was statistically rejected at the 0.05 level by all four of the analyses (Table 3). These results concur with the strong bootstrap and Bayesian values for (Oncorhynchus, Salvelinus) shown in Fig. 2, and they show that the apparently incongruent results from ITS1 do not substantially disrupt the monophyly of these two genera. The relationship (O. tshawytscha, O. kisutch) was statistically supported by both the SH test and the Templeton test using the full nuclear data set (Table 3). By contrast, the SH test on the data set excluding MHC gave a result that was non-signiﬁcant (P ¼ 0:14), and the Templeton test result provided no support for the ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx monophyly of this pair of species. Overall, these ﬁndings are consistent with the moderate (72%) maximum-parsimony bootstrap support for (O. tshawytscha, O. kisutch) in the nuclear data set excluding MHC (Fig. 2D), the strong Bayesian support for this group in this data set, and the 100% bootstrap and Bayesian support for this group from the combined nuclear data set (Fig. 2C). 4. Discussion This is the ﬁrst study of salmonid phylogenetics that uses virtually all of the DNA sequence data currently available. Our analyses of the data from each gene separately, followed by combined analyses of the mitochondrial data, the nuclear data, and the full combined data set, showed that the mitochondrial data yielded levels of resolution and support that were substantially lower than the nuclear data, and that the nuclear data showed higher levels of resolution and support than did the nuclear and mitochondrial data combined (Fig. 2). These ﬁndings indicate that although some of the individual mitochondrial genes provide good evidence for some salmonid relationships (Table 2), the mitochondrial data taken together reduced the strength of the phylogenetic signal. The high noise to signal ratio of the mitochondrial data is probably due to saturation, eﬀects of hybridization, selection (Bernatchez et al., 1995; Wilson and Bernatchez, 1998) or some combination of these processes, and it was not alleviated by removal of third-codon positions for protein-coding genes (results not shown). Such a lack of clear, strong signal in mitochondrial data has probably been responsible for much of the ongoing uncertainty regarding the molecular phylogenetics of Salmonidae. In the combined nuclear DNA data sets, the diﬀerent genes reinforced and complemented one another to a considerable degree, yielding generally well-resolved and well-supported trees (Fig. 2B and C). These trees agree with the results of most previous studies, but also help to resolve some long-standing uncertainties regarding the placement of O. masou, the phylogeny of the Paciﬁc salmon, relationships within Salvelinus, and the sistertaxon to Oncorhynchus. 4.1. Oncorhynchus masou By our combined nuclear phylogenies, Oncorhynchus masou forms a monophyletic group with O. mykiss and O. clarki, and these three taxa comprise the sister-group to the ﬁve other Oncorhynchus species. This result appears to provide a striking case of data from diﬀerent genes complementing and reinforcing one another. Thus, none of the genes analyzed separately provides information on the monophyly of this group of three 15 species, but GH1C, GH2C, ITS1, ITS2, and CYTB each supported the monophyly of (O. mykiss, O. clarki) or (O. mykiss, O. masou). Taken together, the nuclear DNA indicated good support for this clade from maximum-parsimony bootstraps (95% without the MHC data, and 79% with MHC) and Bayesian support values (100% for other data sets). Moreover, the Bayesian analysis of the full nuclear data set also provided 100% support for (O. mykiss, O. clarki), which is consistent with numerous previous studies (e.g., Kitano et al., 1997; McKay et al., 1996; Oakley and Phillips, 1999; Oleinik, 1997; Oohara et al., 1997, 1999; Phillips and Rab, 2001). Results of the SH and Templeton tests (Table 3) are also consistent with the group (O. masou, O. mykiss, O. clarki), although only the SH test on the data set excluding the questionable MHC data achieved statistical signiﬁcance. Previous studies have generally considered O. masou to be basal within the Paciﬁc salmon or within Oncorhynchus as a whole (Kitano et al., 1997; McKay et al., 1996; Murata et al., 1996; Oohara et al., 1997, 1999; see also Oleinik, 2000). Our ﬁndings provide the ﬁrst ﬁrm evidence for its phylogenetic position within the clade of Paciﬁc trout. This inference is consistent with diverse additional forms of evidence from allozymes, morphology, behavior, biogeography, and life history (Table 4), and it should motivate more-detailed evaluation of the evolution of phenotypic traits within this lineage. 4.2. Paciﬁc salmon Our analyses provide a fully resolved and well-supported multi-gene phylogeny for Oncorhynchus excluding O. masou, O. mykiss, and O. clarki. The relationships among O. tshawytscha, O. kisutch, O. nerka, O. gorbuscha, and O. keta shown in Fig. 2B and C have been believed for some time from a variety of morphological, genetic, and other data (Domanico and Phillips, 1995; Domanico et al., 1997; Kitano et al., 1997; McKay et al., 1996; Murata et al., 1993, 1996, 1998; Oakley and Phillips, 1999; Oleinik, 1997; Oohara et al., 1997, 1999; Osinov, 1999; Phillips and Oakley, 1997; Phillips and Pleyte, 1991; ShedÕko et al., 1996; Shedlock et al., 1992; Smith and Stearley, 1989; Takasaki et al., 1994; Thomas and Beckenbach, 1989; Thomas et al., 1986; Utter et al., 1973; Utter and Allendorf, 1994). However, previous analyses have lacked unambiguous or strong support for at least one of the nodes, usually many more. We suspect that the prior lack of conclusive results for the phylogeny of Paciﬁc salmon has been due to a combination of saturation of mitochondrial DNA (e.g., McKay et al., 1996), such that it provides little evidence for more-basal nodes, possible selection on mtDNA (e.g., Bernatchez et al., 1995; Wilson and Bernatchez, 1998), and potential hybridization of O. ARTICLE IN PRESS 16 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Table 4 Evidence from previous studies that (O. masou, O. mykiss, O. clarki) represents a monophyletic group Evidence References (1) Chromosome number of O. masou (2n ¼ 66) is most similar to that of O. mykiss (58–64) and O. clarki (64–68) (2) O. masou similar to O. mykiss in muscle proteins and to O. mykiss and O. clarki in allozymes (3) O. masou is Ômost troutlikeÕ of Oncorhynchus in morphology and behavior Phillips and Rab (2001) Tsuyuki and Roberts (1966), Utter et al. (1973) Neave (1958), Yoshiyasu (1973), Stearley (1992) Tanaka (1965), Tsuyuki and Roberts (1966), Kato (1991), Healey et al. (2001) Chevassus (1979) (4) Some O. masou males and females are iteroparous (5) Male O. masou interbreed best with female O. mykiss, compared to crosses with other salmonids (i.e., low levels of post-zygotic isolation in laboratory studies) (6) O. mykiss, O. clarki, and O. masou have similar life histories, with freshwater residence times 1–2+ years, freshwater populations common (7) O. masou feed and mature during freshwater spawning migration, like O. mykiss and O. clarki but unlike other Paciﬁc salmon (8) Distribution of O. masou is precisely parapatric to that of O. mykiss, with line of demarkation near Amur River, Sea of Othotsk tshawytscha or O. kisutch with one or more of the other three species. Indeed, O. tshawytscha and O. kisutch show a curious tendency to group with O. keta, O. gorbuscha, and O. nerka in analyses of the GH2C, 16S, and ND3 data sets. Given that the fertility of some of the crosses between (O. tshawytscha or O. kisutch) and (O. keta, O. gorbuscha, or O. nerka) is currently high (Chevassus, 1979), it should have been even higher in the past, and hybridization events could have led to the moderate degree of discordance between gene trees observed here (see also Rosenﬁeld et al., 2000). Regardless of such apparent incongruities, the clade (O. tshawytscha, O. kisutch) is strongly supported (99–100% bootstrap or Bayesian support values) by all of the analyses of the full nuclear data set, by the Bayesian analysis of the nuclear data set with MHC excluded, and by the SH and Templeton tests for the full nuclear data set. Rounesfell (1958), Willson (1997) Miller and Brannon (1981), Groot and Margolis (1991) Lee et al. (1980), Kato (1991) 4.3. Salvelinus Our combined nuclear DNA data sets provide a wellresolved and generally well-supported phylogeny for within the genus Salvelinus. This phylogeny is generally concordant with the results of most previous moleculargenetic studies (reviewed in Phillips and Oakley, 1997; Westrich et al., 2002), and also helps in diagnosing some incongruent ﬁndings from single-gene studies. A close relationship between Sv. alpinus and Sv. malma is well supported by morphology (Behnke, 1984; Cavender, 1980), karyotypes (Cavender, 1984; Phillips et al., 1989), allozymes (Crane et al., 1994), and all studies using DNA sequence. Sv. conﬂuentus and Sv. leucomaenis are also usually grouped together by morphology and allozymes (Table 6). However, Sv. conﬂuentus groups strongly with Sv. alpinus and Sv. malma by analyses of mtDNA restriction sites, ND3, ITS2, and satellite DNA Table 5 Summary of evidence from previous studies for a sister-taxon relationship between Oncorhynchus and Salvelinus (rather than Salmo and Oncorhynchus) Evidence References (1) Vitellogenin gene organization groups Oncorhynchus and Salvelinus together (2) Microsatellite gene structure groups Oncorhynchus and Salvelinus together (3) The number of chromosome arms in the karyotype is the same (100) in basal Salvelinus, O. masou, and the Paciﬁc salmon; inferred to have changed to 104 in branch leading to (O. mykiss, O. clarki). (4) Some Salvelinus, O. mykiss, O. clarki, and O. masou have similar life histories, with freshwater residence times 1–2+ years, freshwater populations common (5) Some morphological traits that link Salmo and Oncorhynchus are related to large size and breeding competition, and thus may be convergent (6) Oncorynchus and Salvelinus have diversiﬁed mainly in the Paciﬁc and Nearctic respectively, whereas Salmo is in the Palearctic and Atlantic (7) Oncorhynchus and Salvelinus both diversiﬁed over the same general time period (roughly 6–15 mya), from fossil and molecular-clock evidence Buisine et al. (2002) Angers and Bernatchez (1997) Phillips and Rab (2001) (8) Independent evolution of highly developed anadromy inOncorhynchus and Salmo in diﬀerent ocean basins is not unexpected on ecological grounds Kato (1991), Groot and Margolis (1991), Stearley (1992), Willson (1997) Stearley and Smith (1993) (as reinterpreted here) Angers and Bernatchez (1997) Cavender and Miller (1972), Cavender (1980), Smith et al. (1982), Shedlock et al. (1992), McKay et al. (1996), Oohara et al. (1997) Northcote (1978), Stearley (1992), Hansen and Quinn (1998) ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx 17 Table 6 Evidence from sources other than DNA sequence for relationships among species of Salvelinus, compared to total evidence nuclear-DNA tree inferred here Evidence and reference Phylogeny Morphology (Behnke, 1984) Morphology (Stearley, 1992) Morphology and karyology (Cavender and Kimura, 1989) Satellite DNA (Hartley and Davidson, 1994) Allozymes (Crane et al., 1994) Karyology (Phillips et al., 1989, 2002) This study (Fig. 2B and C) fontinalis, namaycush, ((conﬂuentus, leucomaenis), (alpinus, malma)) ((leucomaenis, fontinalis, (conﬂuentus, namaycush)), (alpinus, malma) (fontinalis, namaycush), ((conﬂuentus, leucomaenis), (alpinus, malma)) leucomaenis, (fontinalis, (namaycush, (alpinus, malma, conﬂuentus))) fontinalis, (namaycush, ((conﬂuentus, leucomaenis), (alpinus, malma))) fontinalis, namaycush, ((conﬂuentus, leucomaenis), (alpinus, malma)) (fontinalis, namaycush), ((conﬂuentus, leucomaenis), (alpinus, malma)) Our nuclear DNA trees were compatible with the results of Phillips et al. (1989, 2002), and identical to the results of Cavender and Kimura (1989). (Grewe et al., 1990; Hartley and Davidson, 1994; Phillips et al., 1994, 1995), but it groups with Sv. leucomaenis by analysis of ITS1 and allozymes (Crane et al., 1994; Phillips et al., 1994). Sv. fontinalis, Sv. namaycush, and Sv. leucomaenis have the same chromosome number, which appears to be primitive within the genus (Phillips et al., 1994); in Sv. fontinalis and Sv. namaycush this karyotype comprises 104 chromosome arms, while Sv. leucomaenis and Sv. conﬂuentus have 100 arms and Sv. malma and Sv. alpina have 98. Sv. fontinalis and Sv. namaycush are basal to the other four species in most previous DNA studies (and allozymes: Crane et al., 1994), though they form strongly supported sister taxa only by the ITS1 analysis of Pleyte et al. (1992). Taken together, DNA-sequence studies of relationships within Salvelinus have yielded strikingly incongruent results, especially with regard to the positions of Sv. conﬂuentus, Sv. fontinalis, and Sv. namaycush. The discordance among phylogenetic studies of Salvelinus appears to be the result of hybridization (Phillips et al., 1994, 1995; Westrich et al., 2002). The main evidence for this hypothesis comes from the many examples of ancient and current hybridization between Salvelinus species. Ancient introgression of mtDNA has been demonstrated for Sv. alpinus and Sv. fontinalis (Bernatchez et al., 1995; Glemet et al., 1998), and for Sv. alpinus and Sv. namaycush (Wilson and Bernatchez, 1998), and ongoing hybridization and introgression have been reported for Sv. alpinus and Sv. namaycush (Wilson and Hebert, 1993), Sv. malma and Sv. conﬂuentus (Baxter et al., 1997), and Sv. conﬂuentus and Sv. fontinalis (Kanda et al., 2002; Redenbach and Taylor, 2002; Spruell et al., 2001). Moreover, most laboratory crosses between Sv. alpinus, Sv. malma, Sv. fontinalis, and Sv. namaycush and other Salvelinus result in fertile oﬀspring (Chevassus, 1979). The extent of hybridization in Salvelinus appears to be higher than within Oncorhynchus (e.g., Allendorf and Leary, 1988; Campton and Utter, 1985; Chevassus, 1979; Dangel et al., 1973; Smith, 1992; Taylor, 2003), though it may be comparable to levels in the coregonids (Ferguson et al., 1978). Introgression of mtDNA in Salvelinus may also in some cases be driven by selection (Bernatchez et al., 1995; Wilson and Bernatchez, 1998), which would tend to amplify its eﬀects. Substantial levels of hybridization throughout the evolutionary history of a group make phylogenetic inference problematic (Arnold, 1992). Indeed, given extensive ongoing hybridization, even the geographic location of Salvelinus samples used for DNA sequencing could substantially aﬀect the inferences. Our combined nuclear data phylogenies, especially those inferred from Bayesian analyses, provide very good resolution and support overall. However, support for the grouping of (Sv. alpinus, Sv. malma, Sv. conﬂuentus, Sv. leucomaenis) appears to derive predominantly from a single gene (ITS2). Similar considerations apply to the support for (Sv. conﬂuentus, Sv. leucomaenis), mainly from ITS1, and (Sv. fontinalis, Sv. namaycush), with support mainly from ITS2. Indeed, when the ITS1 and ITS2 data are excluded, the only group within Salvelinus that is accorded maximum-parsimony bootstrap support over 70% is (Sv. alpinus, Sv. malma) (93%). We believe that the best strategy in such cases is the gathering of DNA-sequence data from as many independently evolving nuclear DNA loci as possible, as well as liberal use of other types of character, such as genome structure, karyotypes and allozymes. Taken together, previous phylogenetic inferences from the use of morphology, allozymes and karyological characters (Table 6) are consistent with our phylogeny, although levels of support for the topologies from such sources of data are diﬃcult to ascertain. Given this concordance among diverse data types, we believe that our DNA phylogeny of Salvelinus (Fig. 2B and C) is very likely to be correct. However, additional data from nuclear genes are needed to rigorously test this hypothesis. 4.4. Sister-group to Oncorhynchus The sister-group to Oncorhynchus has long been believed to be Salmo (e.g., Murata et al., 1996; Phillips and Oakley, 1997; Phillips and Pleyte, 1991; Stearley and Smith, 1993), although analysis of data from the GH1C gene by Oakley and Phillips (1999) provided evidence that Salmo and Oncorhynchus are not sister taxa. Our analyses concur with this result of Oakley and ARTICLE IN PRESS 18 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Phillips, and show that a sister-taxon relationship between Oncorhynchus and Salvelinus is well supported by data from three genes (GH1C, VIT, and ND3), by the combined data sets that exclude mitochondrial DNA, and by SH and Templeton tests using these combined data sets. Evidence for a sister-taxon relationship between Oncorhynchus and Salvelinus also comes from data on gene organization, karyology, life history, morphology, biogeography, and ecology (Table 5). The primary molecular-genetic evidence against (Oncorhynchus, Salvelinus) is the results from analysis of ITS1, which show strong bootstrap support for a sister-taxon relationship between Oncorhynchus and Salmo (Fig. 1D). However, we note that this analysis is compromised by use of an outgroup (H. perryi) that may belong among the ingroup taxa; when the closest available ﬁsh species outside of these genera is used (a cichlid), support for (Oncorhynchus, Salmo) is substantially reduced. Considered together, our analyses, and data from previous studies, constitute strong evidence for a sister-taxon relationship between Oncorhynchus and Salvelinus. These results should compel further evaluation of the data from morphology (e.g., position of the vomerine teeth) that has been used to support a sister-taxon relationship between Oncorhynchus and Salmo. 4.5. Implications for the evolution of salmonid ﬁshes Our results present a number of interesting implications for understanding the evolution of salmonid life history, behavior, and diversiﬁcation. First, the ﬁnding that Salvelinus forms the sister-group to Oncorhynchus indicates that anadromy, in the form of long ocean migrations followed by a return to the natal stream, migration tightly linked to reproduction, and semelparity or a very low degree of iteroparity, has evolved at least twice, once in Salmo and once in Oncorhynchus (Oakley and Phillips, 1999; Stearley, 1992). This parallel evolution of life history also involves large body size for age in both sexes, due to extensive feeding at sea, and strong male– male competition, probably a result of high breeding densities (Crespi and Teo, 2002; Stearley, 1992). Such parallel changes in behavior and life history are ultimately a consequence of the parallel ecological opportunities favoring anadromy in the north Paciﬁc and north Atlantic oceans (Dodson, 1997; Gross et al., 1988; Hansen and Quinn, 1998; McDowall, 1988; Northcote, 1978). Second, the sister-taxa Oncorhynchus and Salvelinus have apparently diversiﬁed in parallel on a large scale: each genus has given rise to exclusively freshwater species (reproductively isolated kokanee forms of O. nerka, Sv. namaycush), forms with interior (freshwater) and sea-run populations (e.g., in O. mykiss and Sv. malma), and exclusively Asian species (O. masou, Sv. leucomaenis) (Rounesfell, 1958; Stearley, 1992; Willson, 1997). These similarities are consistent with the parallel radia- tion of the two genera from a common ancestor, subject to relatively similar selective pressures, opportunities for dispersal, and vicariant events. Third, given that O.masou groups with O. mykiss and O. clarki rather than with the other so-called Paciﬁc salmon, semelparity has apparently evolved twice in Oncorhynchus, once in the lineage leading to O. masou, and once in the lineage leading to (O. tshawytscha, O. kisutch, O. nerka, O. gorbuscha, O. keta). Alternatively, O. masou may be less strictly semelparous than is currently believed, as females of some populations of landlocked O. masou exhibit a small degree of iteroparity (Healey et al., 2001), and one of the main forms of evidence in the literature for semelparity in O. masou appears to have been its presumed phylogenetic position among the Paciﬁc salmon (Kato, 1991). 4.6. Congruence, total evidence, and clade support Our analyses raise a number of issues regarding the use of multiple data sets and criteria for evaluating congruence. First, our results provide good examples of both the strengths and limitations of a conditional combining approach to phylogenetic congruence. Overall, the data sets from the nuclear genes complemented and reinforced one another to yield a robust tree, which is what one hopes that combining of data will achieve (Bull et al., 1993; Cunningham, 1997). By contrast, the results from mtDNA genes tended to contradict the results from nuclear genes, and in the combined data sets the inclusion of mtDNA reduced the degree of resolution and bootstrap or Bayesian support. But because the individual data sets overlapped only partially in the taxa that they included, it was not possible to apply statistically based congruence tests (e.g., Huelsenbeck and Bull, 1996) to our data sets, which could more-objectively justify the exclusion of mtDNA or other data sets such as MHC. Second, the use of bootstrap or Bayesian a posteriori support values to evaluate congruence is subject to important caveats. Majority-rule bootstrap trees may diﬀer from best or strict consensus trees, or the bootstrap proﬁles (sets of bootstrapped trees) from analyses of diﬀerent genes may exhibit little or no overlap (Page, 1996; Sanderson, 1989). Moreover, low bootstraps across a clade can be due to only one or few ÔrogueÕ species whose position is especially uncertain due to long-branch attraction, a paucity of data, or other processes (Page, 1996; Sanderson and Schaﬀer, 2002). For our separate and combined data sets, the bootstrap majority-rule consensus trees were almost always the same as the best or strict consensus trees, subject to the lack of resolution shown in many of the bootstrap trees; these ﬁndings suggest that rogue species are not unduly inﬂuencing our results. Such limitations do not apply to Bayesian a posteriori probability values, which appear ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx to provide a more accurate metric of support for the true tree due to their lack of bias and higher sensitivity to phylogenetic signal (Alfaro et al., 2003). However, such Bayesian probability values are less conservative than bootstraps, and given the vagaries of such factors as sampling error and imprecise model speciﬁcation, Bayesian probability values may in some cases imply strong support for relationships that are accorded only weak support by bootstraps (Alfaro et al., 2003; Douady et al., 2003) or by other metrics such as tree length. In our view, these considerations imply that high Bayesian clade support values should be interpreted cautiously, and that they should be accorded high conﬁdence only in conjunction with high likelihood or parsimony bootstraps, or results from SH or Templeton tests. Third, our analyses of the combined nuclear data sets show that in some cases, adding data from an additional gene can substantially increase maximum-parsimony bootstrap support for some nodes while notably decreasing support for others. Thus, the exclusion of the MHC data from our combined nuclear data set yields a tree with very strong maximum-parsimony bootstrap support for (O. mykiss, O. masou, O. clarki, 95% maximum-parsimony bootstrap) but only moderate support for (O. tshawytscha, O. kisutch, 72%), while its inclusion gives the reverse: weaker support for the former clade (79%), but very high support for the latter (99%). Because MHC is known to be under strong selection in salmonids (Miller and Withler, 1996), we hesitate to include it without reservations, even if when analyzed separately these data provide a tree that does not appear to be unequivocally incongruent with others. In contrast to these parsimony results, the Bayesian a posteriori probabilities for (O. mykiss, O. masou, O. clarki) and for (O. tshawytscha, O. kisutch) remained high (100%) whether or not the MHC data was included (Fig. 2). These results appear to reﬂect the higher sensitivity of Bayesian analysis (vs. maximum parsimony bootstraps) to phylogenetic signal, its increased accuracy in recovering monophyletic groups, and the high susceptibility of maximum parsimony analysis to longbranch attraction, which can erode bootstrap support for the aﬀected clades (Alfaro et al., 2003). Overall, we ﬁnd it diﬃcult to argue against the monophyly of both (O. mykiss, O. masou, O. clarki) and (O. tshawytscha, O. kisutch), as there is considerable support for each clade and no notably conﬂicting results. 4.7. Optimizing future studies One of the most important results of this study is its role in mapping the best route for future molecular-phylogenetic studies of salmonids, with the ultimate goal of a robust tree for all species and subspecies in the family. In our study, the most-informative genes were GH1C, GH2C, VIT, CYTB, ITS1, ITS2, and MHC. Of these 19 genes, evidence for incongruence was observed with the results from ITS1, ITS2, and MHC. Since the apparent cause of the incongruence can be surmised in each case (i.e., hybridization and selection respectively), we believe that the data from these genes should be treated with reservations in combined analyses. Despite such cautions, in each of these cases the apparent incongruence involved the placement of only a few species, the data from the other genes appeared to supercede the problematic eﬀects, and the inclusion of the data from these genes thus increased the robustness of the combined nuclear data tree overall. These results imply that the sequencing of some or all of the genes GH1C, GH2C, VIT, ITS1, ITS2, and possibly CYTB and MHC, for an enlarged set of salmonid species, is likely to provide the best estimate for the phylogeny of Salmonidae, until additional nuclear markers are developed. Indeed, a combined data set that includes only these genes provides almost as well-resolved and supported a phylogeny as the full nuclear data set. When data from the same large suite of taxa are available for a collection of genes, statistically based methods for the analysis of congruence can also be applied, to better assess the extent to which diﬀerent data partitions agree on one true species tree. Such a data set should also allow robust inference of the nature and timing of major events in salmonid diversiﬁcation, using data from fossils (Behnke, 1992; Cavender, 1980; Cavender and Miller, 1972; McPhail, 1997; Smith and Stearley, 1989; Stearley and Smith, 1993), paleoclimatology (Pearcy, 1992), geology (Montgomery, 2000), paleobiogeography (McPhail, 1997; Minckley et al., 1986; Neave, 1958), and the biogeography and ecology of the Esociformes, a freshwater group that has recently been shown to be the sister taxon to salmonids (Ishiguro et al., 2003). In conjunction with recent methods for use of DNA sequence to infer divergence times (Arbogast et al., 2002; Sanderson, 2002), such data will provide a comprehensive, interdisciplinary picture of the adaptive radiation of Salmonidae. Acknowledgments We are grateful to Robert Behnke, Ian Fleming, Andrew Hendry, Todd Oakley, Ruth Phillips, Jerry Smith, Eric Taylor, John Taylor, Kyle Young, and two anonymous reviewers for helpful comments and discussion, and the Natural Sciences and Engineering Research Council of Canada for ﬁnancial support. References Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov ARTICLE IN PRESS 20 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx chain Monte Carlo sampling and bootstrapping in assessing phylogenetic conﬁdence. Mol. Biol. Evol. 20, 255–266. Allendorf, F.W., Leary, R.F., 1988. Conservation and distribution of genetic variation in a polytypic species, the cutthroat trout. Cons. Biol. 2, 170–184. Angers, B., Bernatchez, L., 1997. Complex evolution of a salmonid microsatellite locus and its consequences in inferring allelic divergence from size information. Mol. Biol. Evol. 14, 230–238. Arbogast, B.S., Edwards, S.V., Wakeley, J., Beerli, P., Slowinski, J.B., 2002. Estimating divergence times from molecular data on phylogenetic and population genetic timescales. Ann. Rev. Ecol. Syst. 33, 707–740. Arnold, M.L., 1992. Natural hybridization as an evolutionary process. Ann. Rev. Ecol. Syst. 23, 237–261. Barker, F.K., Lutzoni, F.M., 2002. The utility of the incongruence length diﬀerence test. Syst. Biol. 51, 625–637. Baxter, J.S., Taylor, E.B., Devlin, R.H., Hagen, J., McPhail, J.D., 1997. Evidence for natural hybridization between dolly varden (Salvelinus malma) and bull trout (Salvelinus conﬂuentus) in a northcentral British Columbia watershed. Can. J. Fish. Aquat. Sci. 54, 421–429. Behnke, R.J., 1984. Organizing the diversity of the Arctic char complex. In: Johnson, L., Burns, B.L. (Eds.), Biology of the Arctic Charr. Proceedings of the International Symposium on Arctic Charr. University of Manitoba Press, Winnipeg, pp. 3–21. Behnke, R.J., 1992. Native trout of Western North America. Am. Fish. Soc. Monograph 6, 1–275. Bernatchez, L., Glemet, H., Wilson, C.C., Danzmann, R.G., 1995. Introgression and ﬁxation of Arctic char (Salvelinus alpinus) mitochondrial genome in an allopatric population of brook trout (Salvelinus fontinalis). Can. J. Fish. Aquat. Sci. 52, 179–185. Bull, J.J., Huelsenbeck, J.P., Cunningham, C.C., Swoﬀord, D.P., Waddell, P.J., 1993. Partitioning and combining data in phylogenetic analysis. Syst. Biol. 42, 384–397. Buisine, N., Trichet, V., Wolﬀ, J., 2002. Complex evolution of vitellogenin genes in salmonid ﬁshes. Mol. Genet. Genomics 268, 535–542. Campton, D.E., Utter, F.M., 1985. Natural hybridization between steelhead trout (Salmo gairdneri) and coastal cutthroat trout (Salmo clarki clarki) in two Puget Sound streams. Can. J. Fish. Aquat. Sci. 42, 110–119. Cavender, T.M., 1980. Systematics of Salvelinus from the North Paciﬁc basin. In: Balon, E.K. (Ed.), Charrs: Salmonid Fishes of the Genus Salvelinus. The Hague, Junk, pp. S295–S322. Cavender, T.M., 1984. Cytotaxonomy of North American Salvelinus. In: Johnson, L., Burns, B. (Eds.), Biology of the Arctic Char. Proceedings of the International Symposium on Arctic Char, Winnipeg, Manitoba. Manitoba Press, Winnipeg, pp. 431–445. Cavender, T.M., Kimura, S., 1989. Cytotaxonomy and interrelationships of Paciﬁc basin Salvelinus. Physiol. Ecol. Jpn. Special 1, 49– 68. Cavender, T.M., Miller, R.R., 1972. Smilidonichthys rastrosus, a new Pliocene salmonid ﬁsh. Museum of Natural History, University of Oregon, Bulletin 18. Chevassus, B., 1979. Hybridization in salmonids: results and perspectives. Aquaculture 17, 113–128. Crandall, K.A., Bininda-Emonds, O.R.P., Mace, G.M., Wayne, R.K., 2000. Considering evolutionary processes in conservation biology. Trends Ecol. Evol. 15, 290–295. Crane, P.A., Seeb, L.W., Seeb, J.E., 1994. Genetic relationships among Salvelinus species inferred from allozyme data. Can. J. Fish. Aquat. Sci. 51, 182–197. Crespi, B.J., Teo, R., 2002. Comparative phylogenetic analysis of the evolution of semelparity and life history in salmonid ﬁshes. Evolution 56, 1008–1020. Cunningham, C.W., 1997. Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Empirically testing an iterative procedure for choosing among phylogenetic methods. Syst. Biol. 46, 464–478. Dangel, J.R., Macy, P.T., Withler, F.C. 1973. Annotated bibliography of interspeciﬁc hybridization of ﬁshes of the subfamily Salmoninae. NOAA, Nat. Mar. Fish. Ser. Tech. memorandum NMFS NWFC-1, 48 pp. De Queiroz, A., Donoghue, M.J., Junhyong, K., 1995. Separate versus combined analysis of phylogenetic evidence. Ann. Rev. Ecol. Syst. 26, 657–682. Dodson, J.J., 1997. Fish migration: an evolutionary perspective. In: Godin, J.-G.J. (Ed.), Behavioural Ecology of Teleost Fishes. Oxford University Press, Oxford, UK, pp. 10–36. Domanico, M.J., Phillips, R.B., 1995. Phylogenetic analysis of Paciﬁc salmon (genus Oncorhynchus) based on mitochondrial DNA sequence data. Mol. Phylogenet. Evol. 4, 366–371. Domanico, M.J., Phillips, R.B., Oakley, T.H., 1997. Phylogenetic analysis of Paciﬁc Salmon (genus Oncorhynchus) using nuclear and mitochondrial DNA sequences. Can. J. Fish. Aquat. Sci. 54, 1865– 1872. Douady, C.J., Delsuc, F., Boucher, Y., Doolittle, W.F., Douzery, E.J.P., 2003. Comparison of Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Mol. Biol. Evol. 20, 248–254. Du, S.J., Devlin, R.H., Hew, C.L., 1993. Genomic structure of growth hormone genes in chinook salmon (Oncorhynchus tshawytscha): presence of two functional genes, GH-I and GH-II, and a malespeciﬁc pseudogene, GH-U. DNA Cell Biol. 12, 739–751. Elliott, J.M., 1994. Quantitative Ecology and the Brown Trout. Oxford University Press, Oxford, UK. Ferguson, A., Himberg, K.-J.M., Svardson, G., 1978. Systematics of the Irish pollan (Coregonus pollan Thompson): an electrophoretic comparison with other holarctic Coregoninae. J. Fish Biol. 12, 221–233. Fleming, I.A., 1998. Pattern and variability in the breeding system of Atlantic salmon (Salmo salar), with comparisons to other salmonids. Can. J. Fish. Aquat. Sci. 55 (Suppl. 1), 59–76. Glemet, H., Blier, P., Bernatchez, L., 1998. Geographical extent of Arctic char (Salvelinus alpinus) introgression in brook char populations (S. fontinalis) from eastern Quebec, Canada. Mol. Ecol. 7, 1655–1662. Grewe, P.M., Billington, N., Hebert, D.N., 1990. Phylogenetic relationships among members of Salvelinus inferred from mitochondrial DNA divergence. Can. J. Fish. Aquat. Sci. 47, 984– 991. Groot, C., Margolis, L., 1991. Paciﬁc Salmon Life Histories. University of British Columbia Press, Vancouver, BC. Gross, M.R., Coleman, R.M., McDowall, R.M., 1988. Aquatic productivity and the evolution of diadromous ﬁsh migration. Science 239, 1291–1293. Hall, B.G., 2001. Phylogenetic Trees Made Easy. Sinaeur, Sunderland, MA. Hansen, L.P., Quinn, T.P., 1998. The marine phase of the Atlantic salmon (Salmo salar) life cycle, with comparisons to Paciﬁc salmon. Can. J. Fish. Aquat. Sci. 55 (Suppl. 1), 104–118. Hart, J.L., 1973. Paciﬁc Fishes of Canada. Fisheries Research Board of Canada, Ottawa. Hartley, S.E., Davidson, W.S., 1994. Distribution of satellite DNA sequences isolated from Arctic char, Salvelinus alpinus, in the genus Salvelinus. Can. J. Fish. Aquat. Sci. 51 (Suppl. 1), 277–283. Healey, M., Kline, P., Tsai, C.F., 2001. Saving the endangered Formosa landlocked salmon. Fisheries 26, 6–14. Hendry, A.P., Stearns, S.C., 2003. Evolution Illuminated: Salmon and their Relatives. Oxford University Press, Oxford, UK. Huelsenbeck, J.P., Bull, J.J., 1996. A likelihood ratio test to detect conﬂicting phylogenetic signal. Syst. Biol. 45, 92–98. Huelsenbeck, J.P., Bull, J.J., Cunningham, C.W., 1996. Combining data in phylogenetic analysis. Trends Ecol. Evol. 11, 152–158. ARTICLE IN PRESS B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Huelsenbeck, J.P., Ronquist, F., Nielsen, R., Bollback, J.P., 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294, 2310–2314. Ishiguro, N.B., Miya, M., Nishida, M., 2003. Basal euteleostean relationships: a mitogenomic perspective on the phylogenetic reality of the ‘‘Proacanthopterygii’’. Molec. Phylogenet. Evol. 27, 476–488. Kanda, N., Leary, R.F., Allendorf, F.W., 2002. Evidence of introgressive hybridization between bull trout and brook trout. Trans. Am. Fish. Soc. 131, 772–782. Kato, F., 1991. Life histories of Masu and Amago Salmon (Oncorhynchus masou and Oncorhynchus rhodurus). In: Groot, C., Margolis, L. (Eds.), Paciﬁc Salmon Life Histories. UBC Press, Vancouver, BC, pp. 447–520. Kitano, T., Matsuoka, N., Saitou, N., 1997. Phylogenetic relationship of the genus Oncorhynchus species inferred from nuclear and mitochondrial markers. Genes Genet. Syst. 72, 25–34. Lee, D.S., Gilbert, C.R., Hocutt, C.H., Jenkins, R.E., McAllister, D.E., Stauﬀer Jr., J.R., 1980. Atlas of North American Freshwater Fishes. North Carolina State Museum of Natural History. McDowall, R.M., 1988. Diadromy in Fishes: Migrations Between Freshwater and Marine Environments. Timber Press, Portland, OR. McDowall, R.M., 1997. The evolution of diadromy in ﬁshes (revisited) and its place in phylogenetic analysis. Rev. Fish. Biol. Fish. 7, 443– 462. McKay, S.J., Devlin, R.H., Smith, M.J., 1996. Phylogeny of Paciﬁc salmon and trout based on growth hormone type-2 and mitochondrial NADH dehydrogenase subunit 3 DNA sequences. Can. J. Fish. Aquat. Sci. 53, 1165–1176. McLennan, D.A., 1994. A phylogenetic approach to the study of ﬁsh behavior. Rev. Fish. Biol. Fish. 4, 430–460. McPhail, J.D., 1997. The origin and speciation of Oncorhynchus revisited. In: Stouder, D.J., Bisson, P.A., Naiman, R.J. (Eds.), Paciﬁc Salmon and their Ecosystems: Status and Future Options. Chapman and Hall, New York, pp. 29–38. Miller, K.M., Withler, R.E., 1996. Sequence analysis of a polymorphic Mhc class II gene in Paciﬁc salmon. Immunogenetics 43, 337–351. Miller, R.J., Brannon, E.L., 1981. The origin and development of life history patterns in Paciﬁc salmonids. In: Brannon, E.L., Salo, E.O. (Eds.), Salmon and Trout Migratory Behavior Symposium. University of Washington Press, Seattle, WA, pp. 296–309. Minckley, W.L., Hendrickson, D.A., Bond, C.E., 1986. Geography of western North American freshwater ﬁshes: description and relationships to intracontinental tectonism. In: Hocutt, C.H., Wiley, E.O. (Eds.), Zoogeography of North American Freshwater Fishes. Wiley, New York, pp. 519–614. Montgomery, D.R., 2000. Coevolution of the Paciﬁc salmon and Paciﬁc Rim topography. Geology 28, 1107–1110. Murata, S., Takasaki, N., Okazaki, T., Kobayashi, T., Numachi, K., Chang, K., Okada, N., 1998. Molecular evidence from short interspersed elements (SINEs) that Oncorhynchus masou (cherry salmon) is monophyletic. Can. J. Fish. Aquat. Sci. 55, 1864–1870. Murata, S., Takasaki, N., Saitoh, M., Okada, N., 1993. Determination of the phylogenetic relationships among Paciﬁc Salmonids by using short interspersed elements (SINEs) as temporal landmarks of evolution. Proc. Natl. Acad. Sci. USA 90, 6995–6999. Murata, S., Takasaki, N., Saitoh, M., Tachida, H., Okada, N., 1996. Details of retropositional genome dynamics that provide a rationale for a genetic division: the distinct branching of all the Paciﬁc salmon and trout (Oncorhynchus) from the Atlantic salmon and trout (Salmo). Genetics 142, 915–926. Neave, F., 1958. The origin and speciation of Oncorhynchus. Trans. R. Soc. Canada 52, 25–39. Norden, C.R., 1961. Comparative osteology of representative salmonid ﬁshes, with particular reference to the grayling (Thymallus arcticus) and its phylogeny. J. Fish. Res. Bd. Canada 18, 679–791. 21 Northcote, T.G., 1978. Migratory strategies and production in freshwater ﬁshes. In: Gerking, S.D. (Ed.), Ecology of Freshwater Fish Production. Blackwell Scientiﬁc Publishers, Oxford, UK, pp. 326–359. Oakley, T.H., Phillips, R.B., 1999. Phylogeny of Salmonine ﬁshes based on growth hormone introns: Atlantic (Salmo) and Paciﬁc (Oncorhynchus) Salmon are not sister taxa. Mol. Phylogenet. Evol. 11, 381–393. Oleinik, A.G., 1997. Molecular phylogeny of salmonids: results of an analysis of nuclear and mitochondrial DNA. Genetika 33, 173–177. Oleinik, A.G., 2000. On the rates of evolution of the mitochondrial and nuclear genomes in Salmonid ﬁshes. Biologiya Morya (Vladivostok) 26, 410–416. Oohara, I., Sawano, K., Okazaki, T., 1997. Mitochondrial DNA sequence analysis of the masu salmon––phylogeny in the genus Oncorhynchus. Mol. Phylogenet. Evol. 7, 71–78. Oohara, I., Sawano, K., Okazaki, T., Kobayashi, T., 1999. Reexamination of the molecular phylogeny of the masu salmon of the genus Oncorhynchus. In: Seret, B., Sire, J.-Y. (Eds.), Proc. 5th Indo-Paciﬁc Fish Conference, Noumea, New Caledonia, pp. 417–426. Osinov, A.G., 1999. Salmonid ﬁsh of the genera Salmo, Parasalmo, and Oncorhynchus: genetic divergence, phylogeny and classiﬁcation. J. Ichthyol. 39, 571–587. Page, R.D.M., 1996. On consensus, conﬁdence, and Ôtotal evidenceÕ. Cladistics 12, 83–92. Pearcy, W.G., 1992. Ocean ecology of north Paciﬁc salmonids. University of Washington Press, Seattle, WA. Phillips, R.B., Manley, S.A., Daniels, T.J., 1994. Systematics of the salmonid genus Salvelinus inferred from ribosomal DNA sequences. Can. J. Fish. Aquat. Sci. 51, 198–204. Phillips, R.B., Makoto, M.P., Reed, K.M., 2002. Characterization of charr chromosomes using ﬂuorescence in situ hybridization. Environ. Biol. Fishes 64, 223–228. Phillips, R.B., Oakley, T.H., 1997. Phylogenetic relationships among the Salmoninae based on their nuclear and mitochondrial sequences. In: Kocher, T.D., Sepien, C.A. (Eds.), Molecular Systematics of Fishes. Academic Press, San Diego, CA, pp. 145–162. Phillips, R.B., Pleyte, K.A., 1991. Nuclear DNA and salmonid phylogenetics. J. Fish Biol. 39, 259–275. Phillips, R.B., Pleyte, K.A., Ihssen, P.E., 1989. Patterns of chromosomal nucleolar organizer region (NOR) variation in ﬁshes of the genus Salvelinus. Copeia 1989, 47–53. Phillips, R.B., Rab, P., 2001. Chromosome evolution in the Salmonidae (Pisces): an update. Biol. Rev. 76, 1–25. Phillips, R.B., Sajdak, S.L., Domanico, M.J., 1995. Relationships among charrs based on DNA sequences. Nordic J. Freshwater Res. 71, 378–391. Pleyte, K.A, Duncan, S.D., Phillips, R.B., 1992. Evolutionary relationships of the salmonid ﬁsh genus Salvelinus inferred from DNA sequences of the ﬁrst internal transcribed spacer (ITS 1) of ribosomal DNA. Mol. Phylogenet. Evol. 1, 223–230. Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817–818. Posada, D., Crandall, K.A., 2001. Selecting the best-ﬁt model of nucleotide substitution. Syst. Biol. 50, 580–601. Rannala, B., Yang, Z., 1996. Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. J. Mol. Evol. 43, 304–311. Redenbach, Z., Taylor, E.B., 2002. Evidence for historical introgression along a contact zone between two species of char (Pisces: Salmonidae) in northwestern North America. Evolution 56, 1021– 1035. Regan, C.T., 1914. The systematic arrangement of the ﬁshes of the family Salmonidae. Ann. Mag. Nat. History 13, 405–408. Rosenﬁeld, J.A., Todd, T., Greil, R., 2000. Asymmetric hybridization and introgression between pink salmon and chinook salmon in the Laurentian Great Lakes. Trans. Am. Fish. Soc. 129, 670–679. ARTICLE IN PRESS 22 B.J. Crespi, M.J. Fulton / Molecular Phylogenetics and Evolution xxx (2003) xxx–xxx Rounesfell, G.A., 1958. Anadromy in North American Salmonidae. US Fish Wildlife Ser. Bull. 58, 171–185. Sanderson, M.J., 1989. Conﬁdence limits on phylogenies: the bootstrap revisited. Cladistics 5, 113–129. Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Mol. Biol. Evol. 19, 101–109. Sanderson, M.J., Schaﬀer, H.B., 2002. Troubleshooting molecular phylogenetic analyses. Ann. Rev. Ecol. Syst. 33, 49–72. Scott, W.B., Crossman, E.J., 1973. Freshwater Fishes of Canada. Bull. 184, Fisheries Research Board of Canada, Ottawa. ShedÕko, S.V., Ginatulina, L.K., Parpura, I.Z., Ermolenko, A.V., 1996. Evolutionary and taxonomic relationships among FarEastern salmonid ﬁshes inferred from mitochondrial DNA divergence. J. Fish Biol. 49, 815–829. Shedlock, A.M., Parker, J.D., Crispin, D.A., Pietsch, T.W., Burmer, G.C., 1992. Evolution of the salmonid mitochondrial control region. Mol. Phylogenet. Evol. 1, 179–192. Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of loglikelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116. Slowinski, J., Page, R.D.M., 1999. How should species phylogenies be inferred from sequence data? Syst. Biol. 48, 814–825. Smith, G.R., 1992. Introgression in ﬁshes: signiﬁcance for paleontology, cladistics, and evolutionary rates. Syst. Biol. 41, 41–57. Smith, G.R., Swirydczuk, K., Kimmul, P.G., Wilkinson, B.H., 1982. Fish biostratigraphy of late Miocene to Pleistocene sediments of the western Snake River plain, Idaho. Cenozoic Geol. Idaho, Idaho Bureau Mine Geol. Bull. 26, 519–541. Smith, G.R., Stearley, R.F., 1989. The classiﬁcation and scientiﬁc names of rainbow and cutthroat trouts. Fisheries 14, 4–10. Spruell, P., Bartron, M.L., Kanda, N., Allendorf, F.W., 2001. Detection of hybrids between bull trout (Salvelinus conﬂuentus) and brook trout (Salvelinus fontinalis) using PCR primers complementary to interspersed nuclear elements. Copeia 2001, 1093–1099. Stearley, R.F., 1992. Historical ecology of Salmoninae, with special reference to Oncorhynchus. In: Mayden, R.L. (Ed.), Systematics, Historical Ecology, and North American Freshwater Fishes. Stanford University Press, Stanford, CA, pp. 622–658. Stearley, R.F., Smith, G.R., 1993. Phylogeny of the Paciﬁc trouts and salmons (Oncorhynchus) and genera of the family Salmonidae. Trans. Am. Fish. Soc. 122, 1–33. Sullivan, J., Swoﬀord, D.L., 2001. Should we use model-based methods for phylogenetic inference when we know that assumptions about among-site rate variation and nucleotide substitution pattern are violated? Syst. Biol. 50, 723–729. Swoﬀord, D.L., 2002. PAUP: Phylogenetic Analysis Using Parsimony, version 4.0b10. Illinois Natural History Survey. Champaign, IL. Tanaka, S., 1965. A review of the biological information on masu salmon (Oncorhynchus masou). In: Salmon of the North Paciﬁc Ocean. Part IX. Coho, chinook and masu salmon in oﬀshore waters. Int. North Paciﬁc Fish. Comm. Bull. 16, 75–135. Takasaki, N., Murata, S., Saitoh, M., Kobayashi, T., Park, L., Okada, N., 1994. Species-speciﬁc ampliﬁcation of tRNA-derived short interspersed repetitive elements (SINEs) by retroposition: a process of parasitization of entire genomes during the evolution of salmonids. Proc. Natl. Acad. Sci. USA 91, 10153–10157. Taylor, E.B., 2003. Evolution in mixed company: evolutionary inferences from studies of natural hybridization in Salmonidae. In: Hendry, A., Stearns, S. (Eds.), Evolution Illuminated: Salmon and their Relatives. Oxford University Press, Oxford, UK. Templeton, A.R., 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and apes. Evolution 37, 221–244. Thomas, W.K., Beckenbach, A.T., 1989. Variation in salmonid mitochondrial DNA: evolutionary constraints and mechanisms of substitution. J. Mol. Evol. 29, 233–245. Thomas, W.K., Withler, R.E., Beckenbach, A.T., 1986. Mitochondrial DNA analysis of Paciﬁc salmonid evolution. Can. J. Zool. 64, 1058–1064. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The Clustal X windows interface: ﬂexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl. Acids Res. 25, 4876–4882. Tsuyuki, H., Roberts, E., 1966. Inter-species relationships within the genus Oncorhynchus based on biochemical systematics. J. Fish. Res. Bd. Canada 23, 101–107. Utter, F.M., Allendorf, F.W., Hodgins, H.O., 1973. Genetic variability and relationships in Paciﬁc Salmon and related trout based on protein variations. Syst. Zool. 22, 257–270. Utter, F.M., Allendorf, F.W., 1994. Phylogenetic relationships among species of Oncorhynchus: a consensus view. Conserv. Biol. 8, 864– 867. Westrich, K.M., Konkol, N.R., Matsuoka, M.P., Phillips, R.B., 2002. Interspeciﬁc relationships among charrs based on phylogenetic analysis of nuclear growth hormone intron sequences. Environ. Biol. Fishes 64, 217–222. Wiens, J.J., 1998a. Does adding characters with missing data increase or decrease phylogenetic accuracy? Syst. Biol. 47, 625–640. Wiens, J.J., 1998b. Combining data sets with diﬀerent phylogenetic histories. Syst. Biol. 47, 568–581. Wiens, J.J., Reeder, T.W., 1995. Combining data sets with diﬀerent numbers of taxa for phylogenetic analysis. Syst. Biol. 44, 548– 558. Willson, M.F., 1997. Variation in salmonid life histories: patterns and perspectives. Res. Pap. PNW-RP-498, US Department of Agriculture, Forest Service, Paciﬁc Northwest Research Station, Portland, OR. Wilson, C.C., Bernatchez, L., 1998. The ghost of hybrids past: ﬁxation of Arctic charr (Salvelinus alpinus) mitochondrial DNA in an introgressed population of lake trout (S. namaycush). Mol. Ecol. 7, 127–132. Wilson, C.C., Hebert, P.D.N., 1993. Natural hybridization between Arctic charr (Salvelinus alpinus) and lake trout (S. namaycush) in the Canadian Arctic. Can. J. Fish. Aquat. Sci. 50, 2652–2658. Woram, R.A., Gharbi, K., Sakamoto, T., Hoyheim, B., Holm, L.E., Naish, K., McGowan, C., Ferguson, M.M., Phillips, R.B., Stein, J., Guyomard, R., Cairney, M., Taggart, J.B., Powell, R., Davidson, W., Danzmann, R.G., 2003. Comparative genome analysis of the primary sex-determining locus in salmonid ﬁshes. Genome. Res. 13, 272–280. Yang, Z., Rannala, B., 1997. Bayesian phylogenetic inference using DNA sequences: a Markov Chain Monte Carlo method. Mol. Biol. Evol. 14, 717–724. Yoshiyasu, K., 1973. Starch-gel electrophoresis of hemoglobins of freshwater salmonid ﬁshes in southwest Japan––II. Genus Oncorhynchus (salmon) Bull. Jpn. Soc. Sci. Fish. 39, 97–114.
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project