Inferring the Mode of Speciation From Genomic Data: A Study of the Great Apes (original) (raw)

Journal Article

Department of Ecology and Evolution

, University of Chicago, Chicago, Illinois 60637

Search for other works by this author on:

Department of Ecology and Evolution

, University of Chicago, Chicago, Illinois 60637

Corresponding author: Department of Ecology and Evolution, University of Chicago, 1101 E. 57th St., Chicago, IL 60637. E-mail: ciwu@uchicago.edu

Search for other works by this author on:

Published:

01 January 2005

Navbar Search Filter Mobile Enter search term Search

Abstract

The strictly allopatric model of speciation makes definable predictions on the pattern of divergence, one of which is the uniformity in the divergence time across genomic regions. Using 345 coding and 143 intergenic sequences from the African great apes, we were able to reject the null hypothesis that the divergence time in the coding sequences (CDSs) and intergenic sequences (IGSs) is the same between human and chimpanzee. The conclusion is further supported by the analysis of whole-genome sequences between these species. The difference suggests a prolonged period of genetic exchange during the formation of these two species. Because the analysis should be generally applicable, collecting DNA sequence data from many genomic regions between closely related species should help to settle the debate over the prevalence of the allopatric mode of speciation.

THE allopatric mode of speciation is the tenet of the neo-Darwinian view of speciation (Mayr 1963). In this view, a geographical barrier preventing gene flow is a prerequisite for speciation. Without such barriers, gene exchanges during the process of species formation would obstruct the process as such exchanges would destroy the adaptive gene complexes and obliterate the accumulated differences between nascent species. On the other hand, there is no compelling population genetic reason why divergent adaptation cannot proceed in the presence of continuous gene flow (e.g., Navarro and Barton 2003). A most common mode may be parapatric speciation when nascent species are geographically connected by gene flow (Mayr 1963; Endler 1977). The extreme form of gene flow is represented by sympatric speciation (Dieckmann and Doebeli 1999; Kondrashov and Kondrashov 1999). Parapatric speciation may best be envisioned at the genic level (Wu and Ting 2004) where portions of the genome progressively become divergently adapted and hence nonexchangeable between nascent species. The genealogical history of the genome would therefore be mosaic with disparate divergence time among different loci. While most previous tests of the allopatric vis-à-vis parapatric mode of speciation were based on ecological or biogeographical considerations (Endler 1977; Butlin 1998; Coyne and Price 2000), only a few studies have utilized multiple DNA sequences (Kliman et al. 2000; Machado et al. 2002) for this purpose. This analysis represents a genome-wide perspective on the same issue.

In strict allopatry, all the genes in the genome should have the same divergence history (t in Figure 1A)

Figure 1.—

(A) Allopatric speciation. In strict allopatry, there is no gene flow beyond the time of separation. All genes hence have diverged for a fixed time t and further coalesce with an average length of 2_N_e generations. (B) Parapatric speciation. Under the parapatric model, there is a period of time when gene flow between nascent species is possible. The intensity of shade indicates the strength of the barrier to gene flow. For genomic regions (such as CDSs) associated with reproductive incompatibility, early cessation of gene flow is likely. For regions free of such association (including most IGSs), gene flow may continue until relatively late. (C) Segregation of polymorphisms (m for the ancestral and M for the derived variant) under the allopatric model. The two speciation events, denoted a and b, were separated by _t_′, during which time the effective population size is _N_e′.

but vary in the coalescence time, which is exponentially distributed with mean equal to 2_N_e (_N_e being the effective population size at the time of speciation, Figure 1A). We discuss more complex forms of allopatric speciation later. Note that time is measured in units of generation throughout this report. A large variance in DNA divergence can be due to either variation in t across loci or a larger-than-estimated _N_e, both of which can enhance the variance in divergence among loci. Although there are many studies for estimating _N_e, they all assume constant t across loci, or strict allopatry, precisely what we wish to test. Interestingly, when t is assumed to be a constant, the estimated _N_e's for the ancient species are usually far larger than those for the extant populations (Ruvolo 1997; Takahata and Satta 1997; Chen and Li 2001; Yang 2002; Wall 2003). These studies thus hint at the possibility of nonconstant t.

In this study, we compare coding and intergenic regions for their evolutionary dynamics during speciation. In allopatry, these two types should have the same dynamics but, under the parapatric model of speciation, could have very different histories. Figure 1B illustrates this point, on the assumption that coding sequences are more likely than intergenic sequences to be associated with hybrid incompatibility or differential adaptation. The potential for coding regions to successfully move across nascent species boundaries may be curtailed early (Figure 1B). On the other hand, intergenic regions, experiencing less impediment to their trafficking between nascent species, should continue to be exchangeable until openings in the reproductive barrier are completely sealed. This contrast has been reported for Drosophila between DNA sequences at or near a speciation gene (Ting et al. 2000). A recent report also assumes that the common ancestors of human and chimpanzee went through a period of parapatry (Navarro and Barton 2003). However, the observations were reanalyzed in light of outgroup data and were suggested to result from events unrelated to speciation (Lu et al. 2003; Navarro et al. 2003).

MATERIALS AND METHODS

Sequence data:

We collected 98 common chimpanzee (Pan troglodytes) sequences from the GenBank database, 93 from the 5′-conseusus sequences of Sakate et al. (2003), 19 newly determined full-length cDNA sequences from Ryuichi Sakate and Momoki Hirai (University of Tokyo), and 135 genomic sequences of chimpanzee chromosome 22 corresponding to Ensembl genes of human chromosome 21. Seventy-six gene sequences of gorilla were collected from GenBank. We removed from our analysis MHC sequences, whose genealogy was deeper than human-chimpanzee divergence due to strong balancing selection (Satta et al. 1999). DDBJ/EMBL/GenBank accession numbers of newly determined sequences are AB188273, AB188274, AB188275, AB188276, AB188277, AB188278, AB188279, AB188280, AB188281, AB188282, AB188283, AB188284, AB188285, AB188286, AB188287, AB188288.

The sequences were aligned by using the ClustalW program (Thompson et al. 1994) and corrected by visual inspection. Numbers of synonymous and nonsynonymous substitutions were estimated by the method of Li (1993) with equal weighting among pathways for multiple substitutions in a codon (Pamilo and Bianchi 1993). Fifty-three intergenic sequences were obtained from Chen and Li (2001). Ninety pairs of 2-kb intergenic sequences of human and chimpanzee, which are at least 10 kb apart from genic regions annotated by Ensembl, were obtained from genomic sequences of human chromosome 21 and chimpanzee chromosome 22. Numbers of substitutions of intergenic regions were estimated by using Kimura's two-parameter method (Kimura 1980).

Maximum-likelihood estimate of divergence time and ancestral population size:

We designate ki as the number of synonymous changes for the _i_th sequence [either coding sequence (CDS) or intergenic sequence (IGS)]. The probability of observing ki is given by

\[P\left(k_{i}\right)\ =\ \frac{e^{{-}l_{i}\mathrm{{\tau}}_{i}}}{1\ +\ l_{i}\mathrm{{\theta}}_{i}}\ {{\sum}_{d=0}^{k_{i}}}\frac{\left(l_{i}\mathrm{{\tau}}_{i}\right)^{d}}{d!}\ \left(\frac{l_{i}\mathrm{{\theta}}_{i}}{1\ +\ l_{i}\mathrm{{\theta}}_{i}}\right)^{k_{i}{-}d},\]

where τ_i_ = 2_tui_ and θ_i_ = 4_N_e_ui_ (Equation 5 in Takahata and Satta 1997). li is the length of sequence i and ui is the per-nucleotide substitution rate for the i_th sequence. Equation 1 has two components—the Poisson distribution in the divergence portion (τ_i = 2_tui_) and the “mismatch distribution” in the coalescence portion (θ_i_ = 4_N_e_ui_), where the absence of intragenic recombination is assumed.

Without the outgroup:

We first assume that the substitution rate for CDS (and separately for IGS) is uniform across m loci when the outgroup sequences are not available for calibrating the variation in the mutation rate. Let τ = 2_tu_ and θ = 4_N_e_u_. The log-likelihood for Equation 1 becomes

\[L\left(\mathrm{{\tau},\ {\theta}}\right)\ =\ \mathrm{ln}\ {{\prod}_{i=1}^{m}}\left[\frac{e^{{-}l_{i}\mathrm{{\tau}}}}{1\ +\ l_{i}\mathrm{{\theta}}}\ {{\sum}_{d=0}^{k_{i}}}\frac{\left(l_{i}\mathrm{{\tau}}\right)^{d}}{d!}\ \left(\frac{l_{i}\mathrm{{\theta}}}{1\ +\ l_{i}\mathrm{{\theta}}}\right)^{k_{i}{-}d}\right].\]

The maximum-likelihood estimates (MLEs) of τ and θ were found by numerical iteration.

With the outgroup:

We now use sequences from an outgroup species, say the orangutan, to filter out the variation in ui. Let the divergence time between human and the outgroup be T and the substitution number of the i_th locus between these two species be Ki. We assume that Ki = 2_liTui without considering the coalesence component because, if T is sufficiently large, the impact of 2_N_e should be insignificant. Noting that the divergence time between human and chimpanzee is t, we can now replace τ_i_ and θ_i_ with (t/T)(Ki/li) = α(Ki/li) and (2_N_e/T)(Ki/li) = β(Ki/li), respectively, in Equation 1. α = t/T and β = 2_N_e/T are the two parameters to be estimated by MLE.

The log-likelihood function is

\[L\left(\mathrm{{\alpha},\ {\beta}}|2l_{i}Tu_{i}\ =\ K_{i}\right)\ =\ \mathrm{ln}\ {{\prod}_{i=1}^{m}}\left[\frac{e^{{-}\mathrm{{\alpha}}K_{i}}}{1\ +\ \mathrm{{\beta}}K_{i}}\ {{\sum}_{d=0}^{k_{i}}}\frac{\left(\mathrm{{\alpha}}K_{i}\right)^{d}}{d!}\ \left(\frac{\mathrm{{\beta}}K_{i}}{1\ +\ \mathrm{{\beta}}K_{i}}\right)^{k_{i}{-}d}\right]\]

and the MLE of the two parameters, α and β, can be found by numerical iteration.

Computer simulations:

In the allopatric model, the divergence of all genes is fixed at t (Figure 1A), while in the parapatric model (Figure 1B), the divergence time is uniformly distributed between t and 2_t_. For 1000 loci, random integers in the range of 500–1000 are generated for the number of sites. Coalescent times are generated as exponential distribution whose mean is 2_N_e while _N_e satisfies γ = t/2_N_e = 10. t corresponds to 1% nucleotide divergence with 1.5 × 10−8 substitutions per generation per site. The number of substitutions is assigned according to the Poisson distribution where the mutation rate is uniform among loci. A confidence interval of 95% was calculated by 1000 iterations.

Congruence between gene genealogy and species phylogeny:

Seventy-six gene sequences of human, chimpanzee, gorilla, and orangutan were used for the congruence test. The observed genealogies can be (M, M, m), (m, M, M), or (M, m, M) (see Figure 1C), where m and M are the ancestral and derived variants, respectively. (m, M, M) and (M, m, M), where gorilla shares the derived variant with either chimpanzee or human, are incongruent with the species phylogeny, in which human and chimpanzee are the closest relatives. In our analysis, we first treated each site independently. All CG to TG and CG to CA substitutions were masked from the analysis because of the very high rate of changes at such CpG sites, which often results in genealogies incongruent with the species phylogeny. Adjacent variant sites within the same locus that show the same genealogical pattern are counted as one segment. A locus may have more than one segment showing different phylogenetic patterns, presumably due to recombination.

We designate the observed number of segments that show the pattern of (M, M, m), (m, M, M), and (M, m, M) _a_1, _b_1, and _c_1, respectively, for IGS. Likewise, the numbers are _a_2, _b_2, and _c_2, respectively, for CDS. Under the null hypothesis that _t_′/2_N_e′ is the same between coding and intergenic regions (Figure 1C), the likelihood ratio R is

\[R\ =\ \left(\frac{a_{1}}{n_{1}}\right)^{a_{1}}\ \left(\frac{b_{1}\ +\ c_{1}}{2n_{1}}\right)^{b_{1}+c_{1}}\ \left(\frac{a_{2}}{n_{2}}\right)^{a_{2}}\ \left(\frac{b_{2}\ +\ c_{2}}{2n_{2}}\right)^{b_{2}+c_{2}}/\left(\frac{a_{3}}{n_{3}}\right)^{a_{3}}\ \left(\frac{b_{3}\ +\ c_{3}}{2n_{3}}\right)^{b_{3}+c_{3}},\]

where _a_3 = _a_1 + _a_2, _b_3 = _b_1 + _b_2, _c_3 = _c_1 + _c_2, _n_1 = _a_1 + _b_1 + _c_1, _n_2 = _a_2 + _b_2 + _c_2, and _n_3 = _a_3 + _b_3 + _c_3 (derived from Equation 7 in Wu 1991).

RESULTS AND DISCUSSION

We estimate τ = 2_tu_ and θ = 4_N_e_u_, where u is the per-nucleotide substitution rate, by the maximal-likelihood (ML) method (Takahata and Satta 1997; see materials and methods). Our objective is to test if t is the same between coding and intergenic regions. However, because u many not be the same between two regions, we define γ = τ/θ = t/2_N_e and test if γ is the same between the two regions. γ is the relative divergence accrued after, vis-à-vis before, speciation and should be constant under the null hypothesis of allopatry.

To know how parapatry might affect the estimation of γ when allopatry (i.e., constant t across loci) is incorrectly assumed, we carried out computer simulations. In the simplest case, the divergence time in the allopatric model is fixed at t (Figure 1A), while the divergence time in the parapatric model is uniformly distributed between t and 2_t_ (Figure 1B). More complex simulations have been done but the results can be qualitatively stated as such: parapatry generally results in the underestimation of γ (= t/2_N_e). Even when the true γ is 50% larger in parapatry than in allopatry, as in the case of Figure 1, the estimated numbers are nevertheless in the opposite direction (Table 1)

TABLE 1

Simulation results from the schemes of Figure 1, A (allopatric) and B (parapatric)

| | Divergence timea | γ (expected) | 95% C.I. of γ (estimated) | | | ------------------------------------------------------------ | ------------ | ------------------------- | ------------ | | Model A (allopatric) | t | 10 | 7.36 ∼ 18.54 | | Model B (parapatric) | t ∼ 2_t_ | 15 | 3.48 ∼ 4.72 |

Except for coalescence time.

γ was estimated from 100 rounds of simulations. The parameter values of t, _N_e, and u are given in materials and methods.

TABLE 1

Simulation results from the schemes of Figure 1, A (allopatric) and B (parapatric)

Except for coalescence time.

γ was estimated from 100 rounds of simulations. The parameter values of t, _N_e, and u are given in materials and methods.

. The reason for this seemingly paradoxical result is that, under parapatry, the estimate of 2_N_e is greatly inflated to account for the variation in the level of divergence among loci. Hence, we expect γ to be underestimated when allopatry is incorrectly imposed on data that have a variable divergence time. Coding sequences probably fit this characterization better than intergenic sequences (Figure 1B).

We used 345 CDSs and 143 IGSs from human and chimpanzee and conducted the likelihood-ratio test between the two hypotheses, γCDS = γIGS = γ0 and γCDS ≠ γIGS, where γCDS and γIGS are MLEs for the CDS and IGS, respectively (Takahata and Satta 1997). Under the null hypothesis, the MLE of γ0 is 1.89 and the log-likelihood value is −1098.588 (Table 2)

TABLE 2

Estimation of τ = 2_tu_ and θ = 4_N_e_u_ (see Figure 1A) in pairwise comparisons among human, chimpanzee, and gorilla (γ = t/2_N_e)

| | Human-chimpanzee (_n_C = 345, _n_I = 143) | Human-gorilla (_n_C = 76, _n_I = 53) | Chimpanzee-gorilla (_n_C = 76, _n_I = 53) | | | --------------------------------------------- | -------------------------------------- | ------------------------------------------- | -------- | | H0: γCDS = γIGS = γ0 τCDS | 0.00855 | 0.01299 | 0.01317 | | θCDS | 0.00454 | 0.00500 | 0.00380 | | τIGS | 0.00876 | 0.01094 | 0.01204 | | θIGS | 0.00466 | 0.00421 | 0.00347 | | γ0 | 1.88 | 2.60 | 3.47 | | ln L | −1093.971 | −276.117 | −270.641 | | H1: γCDS ≠ γIGS τCDS | 0.00748 | 0.01286 | 0.01112 | | θCDS | 0.00579 | 0.00514 | 0.00618 | | γCDS | 1.29 | 2.50 | 1.80 | | θIGS | 0.00936 | 0.01099 | 0.01300 | | θIGS | 0.00382 | 0.00414 | 0.00242 | | γIGS | 2.45 | 2.65 | 5.37 | | ln L | −1091.530 | −276.114 | −269.906 | | | P = 0.027 | P = 0.950 | P = 0.224 | |

TABLE 2

Estimation of τ = 2_tu_ and θ = 4_N_e_u_ (see Figure 1A) in pairwise comparisons among human, chimpanzee, and gorilla (γ = t/2_N_e)

. Under the alternative hypothesis, the MLEs for the two regions are γCDS = 1.31 and γIGS = 2.45 and the log-likelihood value is −1096.226 (Table 2). The likelihood-ratio test between the two models yields a significant result (P = 0.027). Because the variation among loci in the number of CpG sites, which exhibit high mutability (Hellmann et al. 2003), may have an impact on our estimation, we reestimated γ by masking all CG to TG and CG to CA substitutions. The likelihood-ratio test leads to the same conclusion (P = 0.006, see supplementary Table 1 at http://www.genetics.org/supplemental/). Strictly speaking, because _N_e may be smaller for the coding than for the intergenic region, as the former is generally less variable than the latter (Pluzhnikov et al. 2002), the null hypothesis should be γIGS ≤ γCDS, making our test conservative. The null hypothesis of γIGS ≤ γCDS is thus rejected.

For the method to be of general use in testing allopatric speciation, the need for DNA sequences should not exceed what we used above. A need for >500 sequences would make the method impractical for most specie pairs. Nevertheless, between human and chimpanzee, 7645 orthologous sequences are available (Clark et al. 2003) to back up the above analysis. For this large dataset, γCDS is 1.20, which leads to an even more significant likelihood ratio (P = 0.0003, see supplementary Table 1). Above 500 sequences, an increase in sample size >500 in this case appears to yield a diminishing return.

To standardize the divergence measure and make it independent of the underlying mutation rate, we also calibrate the human-chimpanzee divergence against the divergence between these two species and an outgroup. We were able to use only 76 CDSs and 53 IGSs from human, chimpanzee, and orangutan for this purpose. It is assumed that the level of divergence between human and orangutan is a function of their divergence time, T, without much influence by the ancestral polymorphism, the contribution of which should be relatively small here. The key parameters are now α = t/T and β = 2_N_e/T (see Figure 1 and materials and methods). By doing so, γ (= α/β) was estimated to be 1.55 and 37.3 for the CDSs and IGSs, respectively. While the estimates are different from those of Table 2 due to both the small sample sizes and the inherent variability in the estimation of γ (see Table 1), the general trend of γIGS ≫ γCDS is observed.

When calibrated against the divergence from the orangutan, the divergence in CDS and IGS between human and chimpanzee can in fact be directly compared since the governing parameters, α = t/T and β = 2_N_e/T, depend only on the common elements, t, T, and 2_N_e. For each locus, we therefore compute the relative divergence _d_R = _d_hc/[(_d_ho + _d_co)/2], where _d_hc, _d_ho, and _d_co are the levels of divergence between human and chimpanzee, human and orangutan, and chimpanzee and orangutan, respectively. The mean of _d_R is 0.522 for CDS and 0.404 for IGS (P = 0.030) while the variance of _d_R is 0.166 for CDS and 0.037 for IGS (P < 10−7). The results suggest that, on average, coding regions have deeper genealogy than intergenic regions and the variation is larger in the former than in the latter, as hypothesized in Figure 1.

The analysis of Table 2 has also been applied to the divergence between gorilla and either human or chimpanzee (node b of Figure 1C). By using 76 coding and 53 intergenic sequences the null hypothesis of allopatry cannot be rejected (P = 0.950 for human-gorilla and P = 0.224 for chimpanzee-gorilla). Although the results are not significant, the chimpanzee-gorilla comparison appears to be very different from the human-gorilla divergence. In the former, γIGS > γCDS and the difference is larger than that in the human-chimpanzee comparison (Table 2). Given the small number of sequences from gorilla, there is little statistical power to resolve the issue at this moment. Nevertheless, chimpanzee and gorilla occupy mainly western Africa, whereas ecological and paleontological evidence suggests proto-humans have migrated to eastern and southern Africa (Leakey et al. 2001). Hence a prolonged period of gene flow between ancestral chimpanzee and gorilla seems plausible.

Finally, we may analyze the joint effect of two speciation events in succession, as shown in Figure 1C. We assume that the species phylogeny of Figure 1C is strictly correct and the two allopatric events are separated by time _t_′ during which the effective population size was _N_e′. The probability of having a genealogy incongruent with the species phylogeny, (m, M, M) or (M, m, M) of Figure 1C, is a function of _t_′/2_N_e′ (Nei 1987; Wu 1991). The null hypothesis, again, is that _t_′/2_N_e′ is the same for coding and intergenic regions. We used 53 intergenic and 76 coding sequences from human (H), chimpanzee (C), gorilla (G), and orangutan (O). Orangutan is used as an outgroup to distinguish the derived mutation, M, from the ancestral state, m. We masked all substitutions at CpG sites and then classified the patterns of independently segregating sites into the three categories shown in Figure 1C.

The proportion of incongruent genealogies is 0.509 and 0.361 for CDS and IGS, respectively (Table 3)

TABLE 3

Number of DNA segments that support any of the three phylogenetic patterns—(HC)(GO), (CG)(HO), or (HG)(CO), where humans (H), chimpanzees (C), and gorillas (G) and orangutans (O) share the variant with one other species only (P = 0.013)

| | (HC)(GO) | (CG)(HO) | (HG)(CO) | | | ---------------------- | ---------- | ---------- | ---------- | | Intergenic (n = 53) | 23 (63.9%) | 6 (16.7%) | 7 (19.4%) | | Coding (n = 76) | 26 (49.1%) | 14 (26.4%) | 13 (24.5%) |

TABLE 3

. The result of the likelihood-ratio test is not significant (P = 0.166), probably due to the small number of sequence fragments. With a larger sample size, say, twice the number of genes in Table 3, this approach should be useful for addressing the issue of allopatric speciation.

By analyzing the divergence among hundreds of DNA sequences, we inferred that the speciation history between human and chimpanzee cannot be the same for coding and intergenic regions. Genomic sequences between closely related species may thus provide new opportunities to settle the debate on the prevalence of allopatric speciation. In a series of analyses, Hey, Wakeley, and colleagues (Wakeley and Hey 1997; Kliman et al. 2000; Machado et al. 2002) addressed the same problem of parapatry using both the polymorphism and divergence data. While their approach utilizes more information per locus, we believe the approach outlined here will be more practical for several reasons. First, in the immediate future, there will be a torrent of data consisting of one sequence per gene per species. Second, polymorphism data will not be useful for resolving the mode of speciation in many species—human vs. chimpanzee being an obvious example. Third, the effect of selection on polymorphism can be more difficult to gauge than that on divergence, making the inference on speciation more difficult.

Finally, allopatric speciation could have more complex patterns than portrayed here. It may happen between deeply subdivided but connected populations where disparate genealogies preexisted when speciation took place allopatrically. Such a model can be seen as a hybrid between parapatry and allopatry. However, if populations can evolve to become differentially adapted and strongly subdivided in the presence of gene flow, it seems plausible that they can continue to diverge without a newly erected geographical barrier to stop gene flow completely. Moreover, the restriction of gene flow imposed by the diverging genomes should continue to strengthen as incompatibilities evolve to encompass larger and larger linkage blocks (Wu and Ting 2004). Testing such a hybrid model may require both the divergence and polymorphism data at the genomic level (Wakeley and Hey 1997; Kliman et al. 2000; Machado et al. 2002). At this moment, testing strict (and simple) allopatry among diverse taxa, as outlined here, seems a logical first step.

Footnotes

Communicating editor: Y.-X. Fu

Acknowledgement

We thank T. Nagylaki, H. Tang, H. Y. Wang, J. Lu, and Y.-X. Fu for providing theoretical advice and/or helping with data analysis; F. C. Chen and W. H. Li for kindly providing the intergenic sequence data; R. Sakate and M. Hirai for the chimpanzee coding sequences; K. Hashimoto and C. K. J. Shen for the macaque cDNA sequences; and J. Shapiro, M. Kohn, B. Harr, M. Long, L. Zhang, I. Boussy, and J. Spofford for comments and discussions.

References

Butlin, R., 1998 What do hybrid zones in general, and the Chorthippus parallelus Zone in particular, tell us about speciation?, pp. 367–378 in Endless Forms: Species and Speciation, edited by D. Howard and S. Berlochers. Oxford University Press, Oxford.

Chen, F.-C., and W.-H. Li,

2001

Genomic divergences between humans and other hominoids and the effective population size of the common ancestor of humans and chimpanzees Am.

J. Hum. Genet.

444

–456.

Clark, A. G., S. Glanowski, R. Nielsen, P. D. Thomas, A. Kejariwal et al.,

2003

Inferring nonneutral evolution from human-chimp-mouse orthologous gene trios.

Science

302

1960

–1963.

Coyne, J., and T. D. Price,

2000

Little evidence for sympatric speciation in island birds.

Evolution

2166

–2171.

Dieckmann, U., and M. Doebeli,

1999

On the origin of species by sympatric speciation.

Nature

400

354

–357.

Endler, J. A.,

1977

Geographic variation, speciation, and clines.

Monogr. Popul. Biol.

–246.

Hellmann, I., S. Zollner, W. Enard, I. Ebersberger, B. Nickel et al.,

2003

Selection on human genes as revealed by comparisons to chimpanzee cDNA.

Genome Res.

831

–837.

Kimura, M.,

1980

A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences.

J. Mol. Evol.

111

–120.

Kliman, R. M., P. Andolfatto, J. A. Coyne, F. Depaulis, M. Kreitman et al.,

2000

The population genetics of the origin and divergence of the Drosophila simulans complex species.

Genetics

156

1913

–1921.

Kondrashov, S., and F. A. Kondrashov,

1999

Interactions among quantitative traits in the course of sympatric speciation.

Nature

400

351

–354.

Leakey, M. G., F. Spoor, F. H. Brown, P. N. Gathogo, C. Kiarie et al.,

2001

New hominin genus from eastern Africa shows diverse middle Pliocene lineages.

Nature

410

433

–440.

Li, W.-H.,

1993

Unbiased estimation of the rates of synonymous and nonsynonymous substitution.

J. Mol. Evol.

–99.

Lu, J., W.-H. Li and C.-I Wu,

2003

Comment on chromosomal speciation and molecular divergence-accelerated evolution in rearranged chromosomes.

Science

302

988

Machado, C. A., R. M. Kliman, J. A. Markert and J. Hey,

2002

Inferring the history of speciation from multilocus DNA sequence data: the case of Drosophila pseudoobscura and close relatives.

Mol. Biol. Evol.

472

–488.

Mayr, E., 1963 Animal Species and Evolution. Belknap Press, Cambridge, MA.

Navarro, A., and N. H. Barton,

2003

Chromosomal speciation and molecular divergence-accelerated evolution in rearranged chromosomes.

Science

300

321

–324.

Navarro, A., T. Marques-Bonet and N. H. Barton,

2003

Response to comment on chromosomal speciation and molecular divergence-accelerated evolution in rearranged chromosomes.

Science

302

988

Nei, M., 1987 Molecular Evolutionary Genetics. Columbia University Press, New York.

Pamilo, P., and N. O. Bianchi,

1993

Evolution of the Zfx and Zfy genes: rates and interdependence between the genes.

Mol. Biol. Evol.

271

–281.

Pluzhnikov, A., A. D. Rienzo and R. R. Hudson,

2002

Inferences about human demography based on multilocus analyses of noncoding sequences.

Genetics

161

1209

–1218.

Ruvolo, M.,

1997

Molecular phylogeny of the hominoids: inferences from multiple independent DNA sequence data sets.

Mol. Biol. Evol.

248

–265.

Sakate, R., N. Osada, M. Hida, S. Sugano, I. Hayasaka et al.,

2003

Analysis of 5′-end sequences of chimpanzee cDNAs.

Genome Res.

1022

–1026.

Satta, Y., H. Kupfermann, Y. J. Li and N. Takahata,

1999

Molecular clock and recombination in primate Mhc genes.

Immunol. Rev.

167

367

–379.

Ting, T., S. C. Tsaur and C.-I Wu,

2000

The phylogeny of closely related species as revealed by the genealogy of a speciation gene, Odysseus.

Proc. Natl. Acad. Sci. USA

5313

–5316.

Takahata, N., and Y. Satta,

1997

Evolution of the primate lineage leading to modern humans: phylogenetic and demographic inferences from DNA sequences.

Proc. Natl. Acad. Sci. USA

4811

–4815.

Thompson, J. D., D. G. Higgins and T. J. Gibson,

1994

CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice.

Nucleic Acids Res.

4673

–4680.

Wakeley, J., and J. Hey,

1997

Estimating ancestral population parameters.

Genetics

145

847

–855.

Wall, J. D.,

2003

Estimating ancestral population sizes and divergence times.

Genetics

163

395

–404.

Wu, C.-I,

1991

Inferences of species phylogeny in relation to segregation of ancient polymorphism.

Genetics

127

429

–435.

Wu, C.-I, and C. T. Ting,

2004

Genes and speciation.

Nat. Rev. Genet.

114

–122.

Yang, Z.,

2002

Likelihood and Bayes estimation of ancestral population sizes in hominoids using data from multiple loci.

Genetics

162

1811

–1823.

Supplementary data

Citations

Views

Altmetric

Metrics

Total Views 388

290 Pageviews

98 PDF Downloads

Since 1/1/2021

Month:	Total Views:
January 2021	1
February 2021	6
March 2021	5
April 2021	9
May 2021	17
June 2021	7
July 2021	8
August 2021	3
September 2021	1
October 2021	2
November 2021	6
December 2021	7
January 2022	7
February 2022	10
March 2022	9
April 2022	7
May 2022	15
June 2022	1
July 2022	7
August 2022	11
September 2022	7
October 2022	2
November 2022	7
December 2022	7
January 2023	5
February 2023	5
March 2023	10
April 2023	5
May 2023	4
June 2023	4
July 2023	8
August 2023	16
September 2023	10
October 2023	5
November 2023	11
December 2023	7
January 2024	10
February 2024	11
March 2024	19
April 2024	26
May 2024	24
June 2024	11
July 2024	12
August 2024	13
September 2024	5
October 2024	5

Citations

71 Web of Science

Inferring the Mode of Speciation From Genomic Data: A Study of the Great Apes (original) (raw)

Abstract

MATERIALS AND METHODS

Sequence data:

Maximum-likelihood estimate of divergence time and ancestral population size:

Without the outgroup:

With the outgroup:

Computer simulations:

Congruence between gene genealogy and species phylogeny:

RESULTS AND DISCUSSION

Footnotes

Acknowledgement

References

Supplementary data

Citations

Views

Altmetric

Email alerts

Citing articles via

Latest

Most Cited

Inferring the Mode of Speciation From Genomic Data: A Study of the Great Apes (original) (raw)

Abstract

MATERIALS AND METHODS

Sequence data:

Maximum-likelihood estimate of divergence time and ancestral population size:

Without the outgroup:

With the outgroup:

Computer simulations:

Congruence between gene genealogy and species phylogeny:

RESULTS AND DISCUSSION

Footnotes

Acknowledgement

References

Supplementary data

Citations

Views

Altmetric

Email alerts

Citing articles via

Latest

Most Read

Most Cited