Seeing through life’s diversity down to the finer points of variation, those that separate one species from another, is no easy task. The cataloging and categorizing of life on Earth, a field called taxonomy (or the more inclusive field of systematics) has been an ongoing human endeavor since ancient times, ranging from Egyptian images of local medically important plant varieties from around 1,500 BC, to the Greek and Roman taxonomists beginning with Aristotle, followed by the early taxonomists of the 16th and 17th centuries (1). However it’s Swedish taxonomist Carl Linnaeus and his 1735 publication Systema Naturae, his first attempt to categorize Earth’s living things and minerals (particularly the 10th edition in 1758 including fauna around the world), who can claim the most credit for the birth of modern taxonomy. During a time of increasing global biological exploration and the associated explosion of “new” plants and animals, Linnaeus cemented the use of binary species names (Genus name followed by species or “trivial name”), solidifying the intellectual and philosophical footing of the biological sciences despite an ongoing atmosphere of confusion and uncertainty among naturalists. While the associated methodology of taxonomy has been refined over time, the binary species name, or Linnean classification, remains the standard in biological nomenclature including for ourselves, Homo sapiens.
The birth of Modern “post-Linnaean” taxonomy spawned a new era of investigations into global biogeography (1). Much progress was made during the 1700’s, including the addition of the rank of Families between Genus and Class to help describe the evolutionary relationships among genera, and the early hints of evolutionary theory such as Lamarkism, invoking the “inheritance of acquired traits”. It wasn’t until 1858 that Charles Darwin and Alfred Russel Wallace, building on these early taxonomists, synthesized the idea of evolution by natural selection, supplanting Lamark’s acquired inheritance theory and revolutionizing our biological worldview. Darwin & Wallace gave us an understanding of how life produces diversity, but also our place in that great diversity, as one branch of a vast and interconnected tree, as opposed to some higher being, inherently and by-design at the top of the pyramid.
Despite the fundamentalist backlash to this rather radical change in our word view, the following decades provided mounting evidence for natural selection’s role in species evolution. With no knowledge of DNA these early evolutionary theorists had little other than observations of collected organisms to work with. Taxonomists spent a great deal of time arguing the best diagnostic characteristics for species or groups, and the various ways to measure these features, the detailed work of morphometrics. Statistical methods, such as discriminant analysis, principle component analysis, and more, were created or adapted from other fields to accurately measure the relationship and variation between such features within and between groupings (2). The Phylogenetic Comparative Method involves using these mathematical relationships to build “phylogenies”, trees of relatedness between samples like a family tree or dog breeding pedigree, and using these to test evolutionary hypotheses, such as suspected relatedness between species or groups. The traditional data for these phylogenies is morphological data, including the dimensions and ratios of fossilized bones, allowing for a broad examination of life’s diversity.
Unfortunately, traditional taxonomic data has it’s limitations (2). Collecting sufficient data to have a representative sample of the population depends on how much variation occurs in the first place, an unknown factor until one starts sampling, and finding enough samples for statistically significant tests may be difficult. Fossil records, generally sparse and incomplete, may struggle to produce enough samples or diagnostic features for good comparison, particularly in creatures that don’t preserve their fossil form well, such as soft bodied invertebrates or fragile snake skeletons. Furthermore, with similar characteristics often produced by convergent evolution rather than shared ancestry (analogy vs homology, which we will discuss below), taxonomists often disagree on which features should be measured to delimit species boundaries.
For example, both dolphins and sharks have fins, however this is doesn’t imply close relatedness as there’s no common, finned ancestor. Dolphins evolved from fin-less mammals much after the sharks. A life in the water means a dorsal foil or rudder (or both) is immensely useful, aiding greatly in stability and maneuvering, thus there is a shared selective pressure for the dorsal surface to evolve an analogous (functionally similar but of different evolutionary origin) fin-like structure, not only in fish and mammals, but also reptiles, amphibians, and a variety of invertebrates. However, all mammals and fish have a spine with vertebrae, a much older homologous (similar in structure and evolutionary origin) feature shared among all the Vertebrata, even the more ancient jawless fishes and rays. Deciding what features to measure or not has been a contentious issue among biologists, and all analyses are dependent on the quality of data going in (crap in, crap out, as the saying goes). Luckily for us, the progress of science relies on not only new data but new techniques for testing hypotheses, and as these new techniques were developed and refined over time (along with data acquisition, as seen in the explosive growth of fossil collections) our ability to make accurate inferences about evolutionary history based on morphology greatly improved.
Traditional morphometrics generally involves the measuring and statistical comparison of some chosen set of features, perhaps the width versus the length of different bones. Modern taxonomists point out that these methods, while useful for the comparison of the features themselves, miss important variance in the overall shape of the organism/s in question. A circle and a square, for instance, may have the same maximum width and height but only one will roll down a 2D hill. In addition to the aforementioned homology versus analogy (or vertebra versus fin) issue, the perceived subjectivity with which an apparent taxonomic authority might choose some characteristic as important, and the varying significance or ‘weighting’ of one character trait over another as decided by different taxonomic experts, was an ongoing problem. As a result, even more rigorous methods were developed, such as the modern field of geometric morphometrics (GMM) (2, 3). Perhaps surprisingly, this newer method relies on the Cartesian coordinate system, developed in the 17th century by French philosopher and mathematician René Descartes. As with many fields, one seemingly unrelated discipline, however far afield in time or space, may provide guiding principles for another.
Rather than measuring features of an organism, GMM uses a collection of shared points known as “landmarks” for each individual in a study (2, 3). In simple 2D analyses, these common landmarks are given their own coordinates on an XY axis, creating an interesting point diagram for each individual in the data set, with which one may play connect-the-dots. The distances between these coordinates gives each individual it’s own configuration of landmarks which can be used without some of the limitations of missing data in traditional methods. What difficulties might occur in not having the exact measurements of a certain bone can often be overcome by knowing multiple shared, interrelated coordinates around that bone, and sampling those instead. Yet even with all these advances, including the refinement of three types of spatial landmarks, semi-landmarks, and developing 3D Cartesian analyses, GMM still faces challenges. The Cartesian method requires all coordinates used in the study are shared between all samples, a difficulty in small/delicate samples, highly similar species, or sketchy fossil records.
There is one characteristic that, despite it’s often highly variable nature, is shared among all living organisms (aside from certain viruses which refuse to behave like the rest of us, but that’s another discussion). Whether simple circular loops in bacteria and mitochondria/plastids, or our much larger nuclear chromosomes and complete genomes, DNA never changes. While the sequence of DNA nucleotides (denoted A,T,G,C) can vary wildly, their individual structure and binding properties remain the same. Additionally, some gene sequences are so vital to important biological roles (e.g. metabolism or chromosome binding) that they’re highly conserved even between humans and the simplest microorganisms, with new mutations heavily punished and eliminated by selection throughout evolutionary history. We share common genes with microorganisms, fungi, plants, insects, fish, snakes, dogs, and of course, with one another. That’s you and I, dear reader, sharing a closer evolutionary relationship and more genetic similarities than we each do with other apes. Despite their similarity, genetic differences accumulate over time as two species diverge down their own evolutionary trajectories, as can be seen in the increasing genetic differences between ourselves and our more distant ape cousins.
DNA thus makes an excellent source of data for phylogenetics (4). Morphological phylogenetics is, in a sense, already making inferences about the genetic relationships between samples. When measuring a characteristic in some species we’re essentially assuming some portion of the variance in morphology is heritable, and that we can base evolutionary relationships on the difference or similarity in this heritable morphological variation. What, then, might we learn if we base our phylogenies directly on DNA, the genetic, heritable, molecular material itself?
This brings us to the field of molecular phylogenetics, an ongoing field within the broader scope of molecular systematics. With the structure of the DNA molecule elucidated, largely with X-ray diffraction data from Rosalind Franklin, assembled and published by Watson & Crick in 1953, the world was introduced to the new and rapidly growing field of molecular biology (4). With the mysteries of life’s chemical functioning being discovered, the field of evolutionary biology also found great value in understanding DNA, RNA and protein sequences. While genome sequencing was some decades off, early molecular systematics work showed hints of what was to come. By the mid 1950s, the technique of electrophoresis, using an electrical current to draw molecules through some resistant medium thereby separating them by size, was being used to compare the size of proteins, or comparing chromosome number and structure. While hardly a quantitative technique, these early methods showed much promise, even matching some well-known morphological relationships while hinting at the possibility of difficulties and unseen variation in others, including a variety of snakes (4, 5).
Of course, many traditional taxonomists weren’t necessarily pleased or accepting of these new molecular methods for exploring species boundaries and relationships (6). Biology of the ’60s and ’70s was somewhat divided, with scientists split between the burgeoning molecular and the established morphological camps, themselves split into factions, during the so called ‘systematist wars’ (4; 6). Despite the embattled hyperbole above, the conflict was real, with traditional schools of thinking reluctant to accept molecular data, defending morphological/functional traits of individual organisms as the ideal subjects of taxonomy. Molecular biologists in return made arguments in favour of molecular markers, initially amino acid variation, as this was all they had available to them . Much support was garnered due to the discrete nature of molecular variation. The one-dimensional, linear, shared nature of of amino acid sequences and their unitary variation allowed molecular biologists to make quantified comparisons of the same proteins from different species (4; 6; 7). The following years saw ongoing advances in molecular techniques, such as the development of DNA-DNA hybridization, and the development of computational methods for building phylogenies using molecular data (4; 8).
With the advent of DNA sequencing methods in the decades following the ’60s, such as Kary Mullis’ now commonly used Polymerase Chain Reaction (PCR) for DNA amplification in early development by 1983 (9), the age of molecular biology was in full swing. Distance-based (DB) methods, based on calculating the pairwise differences between sequences (e.g. the simple uncorrected P-distance is just number of differences divided by the total length of the sequence, excluding any “corrections” for transition ratios and other parameters) to create a matrix of molecular distances which are used to draw a tree, were among the earlier tools used for molecular phylogenetics, however this DB approach can do little more than draw a diagramatic tree from the distance matrix (10; 11). The differences between sequences are reduced to a single distance-value, so any information about the various changes in character states which cause the observed evolutionary relationships, such as their location in the sequence alignment, is lost.
Discrete data methods, including maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (BI) improved on this state of affairs by effectively producing a tree for every column in the alignment, accounting for each specific character change (4; 10; 11). MP searches for the simplest tree possible, one that can account for the sequence data with the minimum number of evolutionary events. ML methods treat a massive, stochastic set of trees and phylogenetic parameters as competing hypotheses and evaluates which best matches the given sequence data. BI, developed by Reverend Robert Bayes in the late 1700’s (another old method co-opted for modern purposes), is also a likelihood approach but involves iteratively, often many billions of times, defining a prior probability function for a hypothesis (tree parameters such as branch length and node placement) which, given a set of information (sequence data), produces a posterior probability function for the competing hypotheses. The higher this posterior score, the more likely that hypothesis is true given the data, thus we can select the most probable tree from the given data (11). These methods require significant computational resources, sometimes taking days occupying a modern PC’s processor to complete billions of alternative phylogenetic trees and likelihood calculations.
Both ML and BI require some defined model of molecular evolution, specifically, of nucleotide/amino acid substitution (4). Correct choice in substitution model is of great importance, as no single model accounts for all mutational rules, particularly when considering the range of molecules and species in question, with varying mutation rates and genomic patterns. An experienced researcher may know which model is most likely to be correct, but subjectivity again lurks nearby such decisions. Standardization of tools such as the Akaike Information Criterion (AIC), jModeltest, and other likelihood approaches can help eliminate the risk by determining statistically which model is best in a reproducible manner and is becoming commonplace (4, 12).
Over time, these improved statistical techniques, such as Maximum Likelihood and Bayesian methods, replaced the earlier distance- or parsimony-based methods as the preferred tool for building phylogenies. Though some MP methods are still valuable, distance methods are more used for their ease and speed for data exploration than for final phylogeny construction like ML and BI (4; 12). Nonetheless, the field of molecular phylogenetics is not without its difficulties. As pointed out by many critics, similar challenges as faced by morphology, such as how to weight characters accurately (in this case, mutations or amino acid substitutions), are common to the fields (4). Further, the rates at which actual mutations occur in different genetic locations and lineages, and difficulties in studying both very closely and very distantly related taxa, continue to plague modern molecular systematists. For instance, the greater the difference between two sequences, the greater the chance that a mutation has occurred twice at the same location. Counting this mutation as one single difference may underestimate the mutation rate or divergence calculations between sequences.
This problem of “multiple hits” is particularly difficult for the faster distance-based methods, though various mathematical tools have been adopted to deal with or at least try to correct or account for such errors (4). Additional challenges include the problem of how to align sequences for analysis, essentially correcting for factors such as read errors, insertions/deletions, gaps, or “domain shuffling”, the transference from one gene to another of a whole segment of a DNA encoding a functional protein domain. Maintaining subjectivity in choosing how to weight character changes, the very thing molecular methods aim to solve, is also an ongoing issue. Various statistical schemes, including improved substitution matrices based on observed probabilities of different substitution rates (such as the Percent Accepted Mutation matrix, or PAM), and the implementation of computer automation have helped remove much of the subjectivity involved in character weighting, sequence alignment, and model selection, though some opponents argue that the fundamental scoring issue remains. Lucky for us, the newer discrete data methods look at characters across multiple samples at once, potentially finding and scoring variation at multiple-hit sites in another taxon, rather than reducing everything to pairwise comparisons, while newer tools, like the aforementioned AIC, help reduce subjectivity and so on (12) . Additionally, these information rich analyses allow for statistical assessment of accuracy, or our level of certainty in accurately representing evolutionary relationships between samples at each branch of the tree.
Despite the challenges, molecular tools are fast becoming a standard in a variety of fields. With increasingly accurate substitution models, faster computing, and better statistical assessment, it’s little wonder that ML and BI are now popular with evolutionary biologists. Genetic diversity, gene flow, population structure; these and more are important metrics available to population geneticists and conservation biologists (13,14). Quantitative genetics, expression profiles, comparative genomic studies, microbial metagenomics; the list of useful molecular tools goes on and on, all variably useful for different fields of research.
Turning to conservation, phylogenetics is particularly useful in identifying so-called “cryptic species”, often closely related species with such similar characteristics that they appear to be the same (14,15,16). Take for example the Death Adders, Acanthophis, an Australo-Papuan genus of short, squat, venomous Elapid snakes, similar in form to vipers but only distantly related. A variety of local species, sub-species, variants and so on have been described from morphology and locale, however the exact number of species and their relationships has been difficult to reconcile. Nonetheless, DNA from similar looking animals can show some significant differences and produce a deeply diverging phylogenetic tree, indicating a longstanding lack of reproduction and gene flow between the them, perhaps even enough to suggest separate species status. This has indeed the case with a number of Australian snakes, including Acanthophis, comprising anywhere from 3 to 7 species depending on one’s morphological opinion. More recent mitochondrial DNA phylogenies showed at least four, perhaps five species (17) with another cryptic species later confirmed from the Kimberly region (16).
Similar work on the well known Mulga or King Brown snake, Pseudechis australis, has also uncovered significant molecular phylogenetic relationships supporting geographically isolated morphological variation discovered in this species. Like the death adders, the complex history of this species’ name demonstrates the difficulties in systematics. Once considered three species, these were all synonymised to the single P. australis based on morphology, and later by now-outdated molecular tools like chromosomal variation and electrophoretic data (18,19). However, newer phylogenies from two P. australis mitochondrial genes published in 2005 showed four deep branches, indicating a significant degree of divergence between the previously identified geographic/morphological groupings (17,20). While the authors “refrain[ed] from assigning names” (20) before formal species descriptions were published, including categorizing morphological, genetic, geographic variation, and submitting holotype references to museum collections and so on, they note the strong evidence for four species, including three small and petite species such as the now well established pygmy mulga (P. weigeli), within this large, powerful, iconic Australian snake. These phylogenies allow us to look at the comparative evolutionary history of Acanthophis and Pseudechis, revealing a dynamic, rather recent history of land-bridging and colonization between Australia and PNG/Indonesia during recent ice ages (17).
More recently, a multi-gene study of the P. australis complex including both mitochondrial and nuclear gene fragments recovered a similar phylogeny further validating the four species (21). Additionally, Bayesian probability methods (specifically, the reversible-jump Markov Chain Monte Carlo, or rjMCMC method) were used to test whether mitochondrial DNA speciation hypotheses were supported by the nuclear DNA data. Lo and behold, the hypothesis of four P. australis species is supported once again, the data capturing both maternal and paternal lines and rejecting the hypothesis of chromosomal recombination as would be seen in a sexually reproducing population.
These are important discoveries, for if what we assume is one large population of a single species is in fact several much smaller ones, our management strategies might underestimate the pressure these small populations are under. Species might very easily be lost, particularly on islands where such cryptic diversity is common. Molecular phylogenetics has uncovered countless new species where we thought there was but one, across a broad range of taxa (15). However, it’s important to remember the shortcoming of these tools, such as requiring more samples per population as samples populations are increased, in developing a rigid study. This requires finding the necessary DNA samples, which can be problematic at best. Further methodological issues arise in what DNA to sample. Is mitochondrial DNA, particularly single “barcoding” genes like COI or ND4, variable enough to accurately represent relationships? If we’re studying populations, might we instead sequence highly variable microsatellite markers throughout the genome? Or should we go all out and sequence the whole genome (14), start to finish, like we’ve done for the human genome and various other interesting or model study organisms?
As always, new methodologies can provide solutions. One obvious answer is to simply sequence more genes, both mitochondrial genes and nuclear genes, to capture variation in both maternal and paternal lines. Many studies successfully “concatenate” or link sequences together (17,21,22), with computational methods dealing with the variable sequence alignment. While including nuclear genes inherited from both parents can show us whether two samples are in fact sharing genes during recent reproduction, the combining and analysis of separate gene sequences, often with variable selection and variation, in a single dataset is rather difficult (21,22,23). While new methods in multi-species coalescent models are able to deal with this issue better than traditional concatenation (21), new technology means other options are also becoming available, such as scanning small parts across the whole genome to categorize variation throughout (14,23,24). These “reduced representation libraries” (24,25) allow for significant variation to be captured in small snapshots to reduce the computational load of a full genome sequencing while also not relying on genes in specific locations.
The rapid rise of RAD-sequencing (Restriction-site Associated DNA polymorphism) since its inception is a sure sign of genome-scan popularity (24). This method digests the whole genome using restriction enzymes, DNA-cutters that slice the DNA strand wherever they find a particular sequence, and sequencing around that site. Briefly, a restriction enzyme like EcoRI cuts wherever it finds GAATTC, hundreds of thousands of times in the human genome. At each cut, a unique identifier, modifier and primer sequence is attached, followed by rounds of PCR to amplify, depending on the method, a sequence of some 100-300bp of genome, each with their own attached identifier sequences to separate out via computational pipeline later on. Thus we get a smattering of a few hundred pairs of DNA sequenced at hundreds of thousands of sites across the whole genome. The phylogenies that can be drawn from such a deep pool of data are likely to capture much of the variation missed by traditional molecular markers, as was shown when Emerson et al. used RAD-seq to explore variation in the Appalacian pitcher plant mosquito, Wyeomyia smithii (26). The programs and processing power needed to sort this pile of code is immense, with methods often relying on cloud-computing to deal with the storage and transfer of information.
As usual, new methods bring new challenges. Aside from the maddening need for increasingly powerful computers, RAD-seq has to contend with the non-randomness of restriction sites, arguably dealt with by using multiple restriction enzymes and increasing the breadth of the dataset again, and biases in GC content, sequencing accuracy and alignment issues (24,25,27). These challenges, as in the past, are being met head on, along with ever decreasing costs and shorter time-frames as automation and efficiency increases. But as these genome scan methods improve, it’s important to not turn our backs on the foundations of our the field. After all, it was morphological variation, often identified from deceased or museum samples, that gave the first hints for species level variation in the Mulga snake and the death adders, clues upon which detailed morphological studies and the more expensive, lab-time consuming molecular work can be based.
So, despite the power of molecular tools, good morphological studies are still themselves invaluable. As late as 2007, the tropical whipsnakes D. olivacea and D. torquata were found to be nine separate species (28). This was based on a thorough, accurate assessment of straightforward morphological and geographical variation in “All locatable specimens of small northern Australian whipsnakes” from ten museums, including the Australian Museum in Sydney and other nation wide collections, the American Museum of Natural History in New York, and the London Natural History Museum. A regression analysis of morphological variation and longitude also showed clear geographical boundaries, as well as sympatry without shared characters between groups or intergrades, as would be expected if these were truly separate. An assessment of geographic and genetic variation would certainly be interesting for evolutionary/ecological reasons and likely confirming but not entirely necessary to rather confidently claim these animals are different species.
Much like the revival of Reverend Bayes and his probabilistic methods for molecular phyogenetics, the maths and cartography of Descartes for modern GMM, or the early foundation of systematics laid down by Carl Linneaus, Lammark, and many, many others, we’d be remiss to turn our backs on older methods simply for the novelty of the modern. With an understanding of the shortcomings and caveats, and clear communication of these to an audience, what we might think of as “old ways” can be incredibly insightful. While many scientific advancements are inarguable improvements on the state of affairs, those in the vanguard may note that such methods can be expensive, time consuming, and even unnecessary to the advancement of our understanding. One need not sequence the genome of a cat and a rat to know they’re of different species, we know enough though morphology, behaviour, biogeography, and so on. With that said, the potential for discovery in mountains of newly available genomic data, as well as plummeting costs, new programs and analytical tools, improved sequencing methods, and reduced timescales from start to finish means that biologists, through phylogenetics, and now phylogenomics, will increasingly look to the very small molecular world for categorizing the great variation of life around us.
References
-
Manktelow, M. (2010) History of taxonomy. Lecture from Dept. of Systematic Biology, Uppsala University. http://atbi.eu/summerschool/files/summerschool/Manktelow_Syllabus.pdf.
-
Ocakoğlu, G., Ercan, İ. (2013) Traditional and Modern Morphometrics: Review. Turkiye Klinikleri J Biostat. 5(1):37-41
-
Richtsmeier, J.T., DeLeon, V.B., Lele, S.R. (2002) The promise of geometric morphometrics. Am J Phys Anthropol; 35:63-91.
-
Suárez-Díaz, E., Anaya-Muñoz, V.H. (2008) History, objectivity, and the construction of molecular phylogenies. Stud Hist Philos Biol Biomed Sci. 39(4):451-68. doi: 10.1016/j.shpsc.2008.09.002
-
Singh, L. (1972) Evolution of Karyotypes in Snakes. Chromosoma (Berl.) 38, 185—236
-
Dietrich, M.R. (1998). Paradox and persuasion: Negotiating the place of molecular evolution within evolutionary biology. Journal of the History of Biology, 31, 85–111
-
Zuckerkandl, E., Pauling, L. (1965) “Evolutionary Divergence and Convergence in Proteins”, in Evolving Genes and Proteins, ed. V. Bryson and H. Vogel (New York: Academic Press, 1965, 97-166)
-
Fitch, W. M., Margohash, E. (1967). Construction of phylogenetic trees. Science, 155(3760), 279–284;
-
Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B., Erlich, H.A. (1988)Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science.;239(4839):487-91
-
Huelsenbeck, J.P. (1995) Performance of Phylogenetic Methods in Simulation. Systematic Biology.Vol. 44, No. 1 (Mar., 1995), pp. 17-48
-
Brown, J.W., Kilmer, A.J. (2003) The state of bayesian phylogenetics: Bayes for the uninitiated.: Kingston: Queen’s University.
-
Posada, D., Buckley, T.R. (2004) Model Selection and Model Averaging in Phylogenetics: Advantages of Akaike Information Criterion and Bayesian Approaches Over Likelihood Ratio Tests. Syst Biol 53 (5): 793-808
-
Ouborg, N.J. (2010) Integrating population genetics and conservation biology in the era of genomics. Biol Lett. 2010 Feb 23; 6(1): 3–6. doi: 10.1098/rsbl.2009.0590
-
Fred W. Allendorf , Paul A. Hohenlohe & Gordon Luikart (2010) Genomics and the future of conservation genetics. Nature Reviews Genetics 11, 697–709 doi:10.1038/nrg2844
-
Bickford, D., Lohman, D.J., Sodhi, N.S., Ng, P.K., Meier, R., Winker, K., Ingram, K.K., Das, I. (2007) Cryptic species as a window on diversity and conservation. Trends Ecol Evol. 2007 Mar;22(3):148-55. Epub 2006 Nov 28.
-
Maddock, S.T., Ellis, R.J., Doughty, P., Smith, L.A., Wüster, W. (2015) A new species of death adder (Acanthophis: Serpentes: Elapidae) from north-western Australia. Zootaxa. 4007(3):301-26. doi: 10.11646/zootaxa.4007.3.1
-
Wüster, W., Dumbrell, A.J., Hay, C., Pook, C.E., Williams, D.J., Fry, B.G. (2005) Snakes across the Strait: trans-Torresian phylogeographic relationships in three genera of Australasian snakes (Serpentes: Elapidae: Acanthophis, Oxyuranus, and Pseudechis). Molecular Phylogenetics and Evolution 34 1–14
-
Mackay, R.D. (1955) A revision of the genus Pseudechis. Proc. R. Zool. Soc. N. S. W. 1953-54: 15-23
-
Mengden, G.A., Shine, R., Moritz, C. (1986) Phylogenetic Relationships within the Australasian Venomous Snakes of the Genus Pseudechis. Herpetologica. 2:2; 215-229
-
Kuch, U., Keough, J.S., Weigel, J. (2005) Phylogeography of Australia’s king brown snake (Pseudechis australis) reveals Pliocene divergence and Pleistocene dispersal of a top predator. Naturwissenschaften (2005) 92:121–127 DOI 10.1007/s00114-004-0602-0
-
Maddock, S.T., Childerstone, A., Fry, B. G, Williams, D.J., Barlow, A., Wüster, W., (2016) Multi-locus phylogeny and species delimitation of Australo-Papuan blacksnakes (Pseudechis Wagler, 1830: Elapidae: Serpentes), Molecular Phylogenetics and Evolution. doi: http://dx.doi.org/10.1016/j.ympev.2016.09.005
-
Jacobsen, F., Friedman, N.R., Omland, K.E. (2010) Congruence between nuclear and mitochondrial DNA: combination of multiple nuclear introns resolves a well-supported phylogeny of New World orioles (Icterus). Mol Phylogenet Evol. 2010 Jul;56(1):419-27. doi: 10.1016/j.ympev.2010.03.035.
-
Blair, C., Murphy, R.W. (2010) Recent trends in molecular phylogenetic analysis: where to next? J Hered.102(1):130-8. doi: 10.1093/jhered/esq092..
-
Davey, J.W., Hohenlohe, P.A., Etter, P.D., Boone, J.Q., Catchen, J.M., Blaxter, M.L. (2011) Genome-wide genetic marker discovery and genotyping using next-generation sequencing. Nature Reviews Genetics 12, 499-510. doi:10.1038/nrg3012.
-
Van Tassell, C.P., Smith, T.P., Matukumalli, L.K., Taylor, J.F., Schnabel, R.D., Lawley, C.T., Haudenschild, C.D., Moore, S.S., Warren, W.C., Sonstegard, T.S. (2008) SNP discovery and allele frequency estimation by deep sequencing of reduced representation libraries. Nature Methods. 5(3):247-52. doi: 10.1038/nmeth.1185.
-
Emerson, K.J., Merz, C.R., Catchen, J.M., Hohenlohe, P.A., Cresko, W.A., Bradshaw, W.E., Holzapfel, C.M. (2010) Resolving postglacial phylogeography using high-throughput sequencing. Proc Natl Acad Sci U S A. 107(37):16196-200. doi: 10.1073/pnas.1006538107
-
Cruaud, A., Gautier, M., Galan, M., Foucaud, J., Sauné, L., Genson, G., Dubois, E., Nidelet, S., Deuve, T., Rasplus, J.Y. (2014) Empirical Assessment of RAD Sequencing for Interspecific Phylogeny. Mol Biol Evol. 31 (5): 1272-1274
-
Shea, G.M., Scanlon, J. D. (2007) Revision of the Small Tropical Whipsnakes Previously Referred to Demansia olivacea (Gray, 1842) and Demansia torquata (Günther, 1862) (Squamata: Elapidae). Records of the Australian Museum. Vol. 59: 117–142. ISSN 0067-1975
Snakes On The Brain 