• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of molbiolevolLink to Publisher's site
Mol Biol Evol. Jan 2012; 29(1): 167–177.
Published online Aug 3, 2011. doi:  10.1093/molbev/msr180
PMCID: PMC3663093

A Limited Role for Gene Duplications in the Evolution of Platypus Venom


Gene duplication followed by adaptive selection is believed to be the primary driver of venom evolution. However, to date, no studies have evaluated the importance of gene duplications for venom evolution using a genomic approach. The availability of a sequenced genome and a venom gland transcriptome for the enigmatic platypus provides a unique opportunity to explore the role that gene duplication plays in venom evolution. Here, we identify gene duplication events and correlate them with expressed transcripts in an in-season venom gland. Gene duplicates (1,508) were identified. These duplicated pairs (421), including genes that have undergone multiple rounds of gene duplications, were expressed in the venom gland. The majority of these genes are involved in metabolism and protein synthesis not toxin functions. Twelve secretory genes including serine proteases, metalloproteinases, and protease inhibitors likely to produce symptoms of envenomation such as vasodilation and pain were detected. Only 16 of 107 platypus genes with high similarity to known toxins evolved through gene duplication. Platypus venom C-type natriuretic peptides and nerve growth factor do not possess lineage-specific gene duplicates. Extensive duplications, believed to increase the potency of toxic content and promote toxin diversification, were not found. This is the first study to take a genome-wide approach in order to examine the impact of gene duplication on venom evolution. Our findings support the idea that adaptive selection acts on gene duplicates to drive the independent evolution and functional diversification of similar venom genes in venomous species. However, gene duplications alone do not explain the “venome” of the platypus. Other mechanisms, such as alternative splicing and mutation, may be important in venom innovation.

Keywords: gene duplications, venom, platypus, evolution


Platypuses are venomous mammals. Male platypuses possess spurs on their hind legs that deliver venom during the breeding season. In humans, envenomation causes both excruciating immediate and long-lasting pain that does not respond to traditional painkillers such as morphine. Other symptoms of envenomation include edema, a drop in blood pressure, and blood coagulation (Martin and Tidswell 1895; Fenner et al. 1992). Proteomic studies in the 1990s revealed that platypus venom contains at least 19 different peptide components (de Plater et al. 1995) of which only three fractions were characterized: C-type natriuretic peptides, defensin-like peptides, and nerve growth factors. More recently, a molecular transcriptomic study of an in-season venom gland led to the identification of proteases, spider venom alpha-latrotoxin-like peptides, cysteine-rich secretory proteins, cytolytic toxin-like peptides, and venom proteins with homology to stonefish stonustoxin (Whittington et al. 2010).

Similar venom toxin gene families are found in a diverse range of venomous animals including shrews, snakes and other reptiles, amphibians, fish, mollusks, insects, cnidarians, and echinoderms. These toxins have evolved in a convergent manner in different lineages independently (Fry et al. 2009). Protein scaffolds that are recruited into venom peptides include chitinase, cystatin, defensin, hyaluronidase, Kunitz, lectin, lipocalin, natriuretic peptide, peptidase S1, phospholipase A2, sphingomyelinase D, SPRY, and cysteine-rich secretory proteins including the AVIT family (Fry et al. 2009).

Gene duplications appear to play an important role in the recruitment and diversification of toxin genes (Kordis and Gubensek 2000; Reza et al. 2006; Lynch 2007; Juárez et al. 2008). The evolutionary fate of duplicated genes was first developed by Ohno (1970). He proposed that immediately following a duplication event, one gene copy is expected to maintain the “ancestral” role, whereas the other copy is free to evolve new function (neofunctionalization). It is now accepted that duplicated genes can be also maintained through the sharing of ancestral gene function (subfunctionalization) and that selective pressures can play a role in the preservation of a gene duplicate (Bergthorsson et al. 2007). Across frog species, caeruleins, which are skin toxin decapeptides, have arisen by independent duplications (Roelants et al. 2010). Venom serine proteases in the North American shrew have evolved adaptively through amino acid changes to regulatory loops surrounding the active site, also following gene duplications (Aminetzach et al. 2009). Further examples of gene duplications in venom include contoxins in the cone snail (Duda and Palumbi 1999), a multitude of gene families including phospholipase A2 (Gutiérrez and Lomonte 1995), serine proteases (Kini 2005), serine protease inhibitors (Zupunski et al. 2003), C-type lectins (Ogawa et al. 2005), coagulation factor V (Minh Le et al. 2005; Reza et al. 2006), and three finger toxins (Fry et al. 2003), across a range of snake species and cysteine-enriched toxins in scorpions (Zhijian et al. 2006).

Venom toxins across animal taxa are members of some of the fastest duplicating gene families in humans: serine proteases, defensins, and natural killer cell (NKC) receptor genes (lectins) (Bailey et al. 2002). In humans, these genes are involved in pathogen defense. In platypus, antimicrobial peptides have given rise through tandem gene duplications to venom peptides (Whittington et al. 2008). The rapid duplication rate in immune genes, driven by intense competition between host and pathogen, is likely to be favorable toward the initial creation of new traits, such as venom function. This is then followed by subsequent duplications that are likely to reinforce the venom system. Venom genes, like immune genes, are also likely to face strong adaptive pressures. In most venomous animals, such pressure is brought about through predator–prey interactions where venom is used to immobilize prey and defend against predators and/or circumvent prey resistance (Poran et al. 1987; Daltry et al. 1996; Biardi et al. 2006; Calvete et al. 2009). However, in platypus, this adaptive pressure is most likely exerted by intraspecific competition, to assert dominance over other males during the breeding season (Grant 2007).

Prior studies on the selective constraints of gene duplicates in venom have focused on the venomous snakes and the cone snail. In snakes, venom gene families such as phospholipase A2, serine protease, C-type lectin-like proteins, metalloproteinases, and serine proteinase inhibitors appear to have evolved through neofunctionalization under strong selective constraints (Kordis and Gubensek 2000; Lynch 2007; Juárez et al. 2008; Aminetzach et al. 2009; Casewell, Wagstaff, Harrison, Renjifo, et al. 2011). In the venomous gastropod genus Conus, a number of contoxin families block sodium and calcium channels and neural receptors in prey species and have diversified their gene function through strong positive selection following extensive duplications (Duda and Palumbi 1999). In addition to toxin peptides that have evolved through gene duplications, genes involved in the biosynthesis of toxins, such as the sxt gene cluster involved in the production of cyanobacteria neurotoxins, have also undergone duplications followed by adaptive evolution (Murray et al. 2011).

Although previous studies suggest that duplicated genes are the major driver of venom evolution, the direct impact of gene duplications on the evolution of venom has not been characterized. This is primarily due to difficulties in assigning gene duplicates in the absence of a sequenced genome. Given that the platypus is the only venomous animal with a sequenced genome (Warren et al. 2008) and venom gland transcriptome (Whittington et al. 2010), we have taken a novel genome-wide approach to assess how duplications have contributed to venom function. To place the role of gene duplication in the context of platypus venom innovation, here, we identify lineage-specific gene duplicates in the platypus genome that are also expressed in an in-season venom gland. Our results suggest that gene duplications do not appear to be the major driver of toxin evolution in the platypus.

Materials and Methods

Identification of Platypus-Specific Gene Duplications

We used homology data from Ensembl Compara Release 61 to identify all platypus genes that are duplicated in the monotreme lineage including genes that are derived from multiple lineage-specific duplications (Vilella et al. 2009; Flicek et al. 2010). All cDNA and protein sequences from monotreme gene duplicates were extracted. We took into consideration the fact that gene loss in other lineages will lead to incorrect inferences of monotreme duplications. A phylogenetic tree may appear to show a species-specific duplication when, in fact, the topology is the result of multiple gene losses in other species (Casewell, Wagstaff, Harrison, and Wüster 2011). However, due to the large number of species in Ensembl, we expect this type of error to be negligible. A list of nonredundant duplicated gene pairs containing monotreme-specific duplication events was constructed (supplementary file 1, Supplementary Material online). Each pair contained at least one gene expressed in the venom gland. We also constructed expansion groups comprising two or more genes where at least one gene is expressed in the venom gland. To confirm Ensembl gene predictions in the absence of expression in the venom gland, we checked for the expression of these genes in other tissue transcriptomes.

Identification of Gene Duplicates Expressed in the Venom Gland

A normalized cDNA library from a breeding season platypus venom gland was sequenced using the Illumina platform, as described in Whittington et al. 2010. The raw data set comprised 19,069,168 reads of 36 nucleotides in length. These reads were mapped against the platypus genome assembly (5.0) using MAQ (Li), and the number of reads matching each Ensembl platypus gene was tallied and used as digital expression values (see Whittington et al. 2010). Multimapped reads were treated in a uniform random manner to increase transcriptomic coverage and decrease false negatives in the detection of expressed duplicates. Only monotreme genes with read counts above 100 were used for further analysis. Protein searches were conducted using BLASTP (Altschul et al. 1997) against Tox-Prot, a UniProtKB/Swiss-Prot database of animal toxins (Release 2010_10) (Jungo and Bairoch 2005) with an E value cutoff of 1 × 105.

Identification of Secreted Gene Duplicates

To date, all known toxin peptides are secretory proteins (Fry et al. 2009). We used SignalP 3.0 hidden Markov models (Bendtsen et al. 2004) and Gene Ontology (GO) (Ashburner et al. 2000) terms assigned by basic local alignment search tool (BLAST) to predict the cellular localization of expressed protein sequences. Disulfide cross-linking, common to secreted proteins, enhances stable tertiary structures (Mamathambika and Bardwell 2008). To confirm that our predictions are secreted, we compared the ratio of cysteine residues with other amino acid residues in secretory and nonsecretory peptides. Functional annotations were inferred by sequence similarity to human proteins. However, we note that as these genes are the result of gene duplication, it is possible that completely novel function has evolved.

Test for Adaptive Selection

We used PAML (Yang 2007) to obtain dn/ds values for each pair of duplicates under the pairwise M0 model. Since an averaged omega ratio (M0) over the entire length of a gene is not a sensitive measure of positive selection, which may only occur on a small subset of sites (Anisimova et al. 2001), we tested branches of venom gland expressed genes for adaptive selection using the branch-site test, “test 2.” We note that both these approaches give conservative estimations of positive selection. Branch lengths taken from M0 model runs were used as initial branch lengths in the branch-site test. Pairs where branch lengths were greater than five nucleotide substitutions per codon were removed from further analysis. Alignments were constructed using PRANK and a codon substitution matrix (Löytynoja and Goldman 2008). Two test 2 runs were conducted. Sequences showing convergence of likelihood values with a difference of <0.001 were kept. Alignments were carefully checked to ensure gene duplicates were not incorrectly assigned due to short contig lengths. The chi-square distribution with one degree of freedom was used to evaluate the likelihood ratio test statistic for positive selection. FDR was used to adjust P values for multiple testing (Hochberg and Benjamini 1990). We checked for significant (P < 0.05) cases of gene conversions using GENECONV (Sawyer 1999).

Phylogenetic Trees

Evolutionary relationships of protease protein sequences were constructed using MEGA5 (Tamura et al. 2007). The bootstrapped (500 replicates) trees were inferred using the neighbor-joining (NJ) method with evolutionary distances computed using the Poisson correction method. Ambiguous positions were removed for each sequence pair.


Identification of Duplicated Gene Groups

Fifty-six percent (10,676,863 reads) of the raw reads were aligned to the platypus genome. We identified 378 pairs of monotreme-specific duplicates, where at least one gene copy (>100 Illumina reads) was one of the 5,410 protein-coding genes expressed in the platypus venom gland. Of these, 327 showed venom gland expression in only one duplicated copy, which is indicative of venom-specific function evolution following gene duplication. A summary of the number of monotreme-specific duplicates is shown in table 1.

Table 1.
Number of Monotreme-Specific Duplications.

Gene family expansions were also identified. Forty-two gene duplication events leading to a monophyletic clade containing three or more genes were identified. In each case, a gene duplication occurred within an internal, or ancestral, node. In the majority of cases, venom gland expression was not detected for all members of a clade. After removing genes on short contigs, we found that the maximum number of genes expressed in the venom gland in any monotreme-specific clade was three. We found evidence of expression of 34% of gene duplicates that were not found in in-season venom gland in other tissue transcriptomes.

Duplicated Venom Genes

The vast majority of genes expressed in the venom gland were not significantly similar to any known toxin proteins. Only 107 genes (from 5,410 genes) matched existing toxin and venom proteins from the Tox-Prot database (E value < 1 × 105). Of these, only 16 genes were monotreme lineage-specific duplicates (supplementary file 2, Supplementary Material online). Only two genes of these possessed signal peptides (ENSOANG00000001418, ENSOANG00000012443).

To explore the possibility that unique toxin genes exist in the platypus, it is necessary to distinguish likely toxin peptides from genes with other functional roles that are also expressed in the venom gland. We focused on expressed genes that have duplicated in the monotreme lineage and had features of secretory peptides, as we would expect that these are more likely to have toxin function (Fry et al. 2009). Forty-two putative secretory genes were identified using a predictive algorithm implemented in SignalP and GO term analysis based on similarity to human secretory proteins (table 2). GO terms were used due to the fragmentation of the platypus assembly, and we recognize that some secreted genes may have truncated 5′ termini, resulting in missing signal motifs and nondetection using SignalP.

Table 2.
List of Duplicated Putative Platypus Venom Genes that Are Expressed in an In-Season Venom Gland and Are Likely to be Secreted.

We expected that the cysteine contents of putative secretory genes would be higher than average compared with intracellular genes due to disulfide bonding which provide structural stability (Clark 2005). The average ratios of cysteine versus noncysteine residues for secreted and nonsecreted genes were 1:52 (standard error [SE] = 7) and 1:77 (SE = 4), respectively.

Twelve of these genes share similarity with known venom peptides in other species. These included five genes involved in protein degradation (three serine proteases and two metalloproteinase), two protease inhibitors, two C-type lectin genes, an isomerase, and one SPRY domain-containing gene similar to snake Ohanin protein.

Adaptive Selection

Twenty of 159 pairs of duplicates (~12%) show evidence of positive selection (FDR < 0.05) as identified using the branch-site model implemented in PAML (test 2) (table 3). To eliminate errors in our results, we took a series of steps to ensure accurate inference of gene duplication and accurate alignment. We filtered pairs based on gene lengths, sequence divergence, and likelihood of past gene conversions in order to eliminate factors that can lead to incorrect results. First, we removed gene duplicates that did not show overlap in multiple alignments with other species. We then aligned sequences using PRANK incorporating a codon substitution model. This produces alignments with a lower number of false positive results compared with other multiple aligners (Fletcher and Yang 2010). Due to the fragmentation of the platypus genome assembly, many genes are truncated. This could lead to incorrect assignment of gene duplicate, where one gene, split over two or more contigs, is considered to be two or more genes and treated as gene duplicates. Truncated genes can also lead to gene prediction errors, which may elevate the level of positive selection. To overcome these issues, we only kept putative gene duplicates that overlapped by 60 bp or more when aligned. Short uninformative alignments were removed from the selection analysis. Highly divergent sequences can lead to spuriously high levels of positive selection due to alignment errors, different codon usage, and different nucleotide composition (Fletcher and Yang 2010). To address this, we filtered gene pairs based on tree branch lengths. We further removed all gene pairs that showed signs of gene conversion, as recombination events can also lead to elevated rates of adaptive selection (Anisimova et al. 2003).

Table 3.
Venom Gland-Expressed Monotreme-Specific Duplicate Pairs under Positive Selection.

We also performed selection tests on duplication groups (>2 genes) that are likely to have toxin roles. We identified strong positive selection acting on the lineage of an expressed protease gene (P << 0.001; ENSOANG00000002931) belonging to a group of four serine protease inhibitors that have expanded in the monotreme lineage. All genes contained the Kunitz-type domain, which has been identified in a wide range of venomous species, including snake, wasp, scorpion, sea anemones, and coral (Fry et al. 2009). The branch of an expressed C-type lectin gene (ENSOANG00000019797), within a monotreme group containing six genes, was also under positive selection P < 0.05.


Our results show that only a small proportion of putative venom genes are duplicated with a smaller subset displaying evidence of adaptive evolution. Here, we discuss the role of gene duplications in platypus venom evolution in secreted proteases, protease inhibitors, C-type lectins, ion channel proteins, isomerases, and beta-2-microglobulin (B2M).


Whittington et al. (2010) identified 14 serine proteases in the platypus genome. Only three of the 14 appear to result from lineage-specific gene duplications, including two genes which derive from a lineage-specific clade-expansion of five genes that share similarity with human trypase (fig. 1) and one gene that is most similar to human chymase. A platypus peptidase with high similarity to a metalloprotease, endoplasmic reticulum aminopeptidase 1 (ERAP1) was also identified. Both ERAP1 and chymase are involved in the regulation of blood pressure in humans. The physiological function of ERAP1 in humans suggests that it may cause hypotension by cleaving angiotensin II (Hattori et al. 2000). Hence, it is possible that this gene duplicate may contribute to the hypotensive effect of platypus envenomation (Martin and Tidswell 1895). On the other hand, chymase, a member of the peptidase S1 family, cleaves inactive angiotensin I into angiotensin II, causing an increase in blood pressure (Wu et al. 2005) which is not consistent with the physiological effect of platypus envenomation. It is possible that this hypertensive effect may be neutralized by the expression of ten serine protease inhibitors, which are found in venom (Whittington et al. 2010). The presence of both hypertensive and hypotensive proteins may aid in the regulation of blood pressure inside the venom gland. Additionally, proteases typically possess a wide range of functions, including smooth muscle contraction, inflammation, blood clotting, degradation of extracellular matrix, and pain (Wu et al. 2005). Therefore, it is possible that the gene duplicate was not selected for its role in hypertension but rather for other functional properties. In support of this, a platypus chymase-like duplicate also shares high levels of homology to a toxin from short-trailed shrew (BLAST; Accession: Q76B45; E value = 5 × 1031), which has kallikrein-like activity and dilatory effects on blood vessel walls (Kita et al. 2004).

FIG. 1.
Evolutionary relationship of serine protease protein sequences. There were 366 positions in the final data set. Bootstrap values greater than 70% are shown. Tree is drawn to scale based on the number of amino acid substitutions per site. Platypus sequences ...

Homology of platypus peptidases to chymase and trypase suggests that platypus venom serine proteases arose from different ancestral genes to snake peptidases, which are derived from tissue kallikrein (fig. 1) (Deshimaru et al. 1996). Venom peptidases have also independently evolved in shrew, reptiles, and insects (Fry et al. 2009).

Protease Inhibitors

Two pairs of serine protease inhibitors have resulted from lineage-specific gene duplications: SERPINA3 and A2M. Both genes are associated with acute phase response in humans. Serpins belong to a superfamily of protease inhibitors that are found in gene clusters and evolve through a series of tandem gene duplications and control coagulation pathways and immune processes (van Gent et al. 2003). SERPINA3 inhibits chymase, which has also experienced a gene duplication in monotremes, suggestive of coevolution of the receptor and ligand pair in venom function. Serpin inhibition of chymase leads to in a drop in blood pressure (Rubin et al. 1990), consistent with observations in rabbits that have been injected with platypus venom. The platypus gene duplicate may also have a protective function: In wasp venom, serpins are believed to protect host tissues from toxin effects generated by serine proteases (Colinet et al. 2009). A2M is also a protease inhibitor that is capable of eliminating proteolytic effects of several crotalid snake venoms (Kress and Catanese 1981). A2M is homologous to C3, a thiolester-containing protein of the vertebrate complement and homolog of cobra venom factor (Sottrup-Jensen et al. 1985), suggesting a convergent role of A2M-like domains in venom. This is the first time, to our knowledge, that two A2M genes have been identified in the venom component of any species. We hypothesize that it functions in a protective and regulatory capacity.

Ion Channel Protein

A duplication of an ion channel gene similar to TRPV6 was identified. This member of the transient receptor potential (TRP) ion channel family plays a role in neurotransmission, muscle contraction, and exocytosis using calcium as a secondary messenger in humans (Clapham et al. 2001). Given its expression in the venom gland, this receptor may function in the secretion of venom.


We identified a secreted gene duplicate belonging to the peptidylprolyl cis-trans isomerase (PPIase) family. PPIases catalyze the cis-trans isomerization of prolines. Alterations in the folding of proteins can lead to novel structures and toxin function. This is supported by the discovery of a platypus venom ld-isomerase, which may protect toxins from protease degradation and increase toxin potency (Torres et al. 2006; Torres et al. 2007). In addition to cellular signaling (Luban et al. 1993) and transcriptional regulation (Rycyzyn and Clevenger 2002), PPIases in cone snail have been documented to accelerate the folding of disulfide-rich toxins containing prolines by acting as chaperons in protein folding (Safavi-Hemami et al. 2010). The PPIase may possess a similar role in platypus venom.


An expressed B2M duplicate was indentified. It is the third most highly expressed gene in the Illumina data set, higher than any of the characterized venom peptides. Expression of B2M in venom or venom tissues in other venomous species has not been described, but its high expression level and lineage-specific duplication warrants further investigation. Due to the absence of reports of B2M in the venom of other species, we checked whether its high level of expression in platypus was due to the presence of noncoding RNA. We mapped reads back to the genome and confirmed that its expression is indeed due to the protein-coding gene itself. High levels of B2M in plasma cause destructive osteoarthropathies in humans, including joint pain and swelling (Drüeke 1999). These symptoms are a result of B2M forming nonselective voltage independent ion channels or pores in phospholipid bilayer membranes (Hirakura and Kagan 2001). Formation of such channels could compromise membrane potential and allow toxic extracellular components, such as calcium, into the tissue of an envenomated victim, while causing the loss of crucial ions including potassium and magnesium (Hirakura and Kagan 2001). It may also cause cell death through the destruction of intracellular structures and leakage of lysosomal enzymes (Hirakura and Kagan 2001).

C-Type Lectins

Two of 12 C-type lectin lineage-specific duplicates were expressed in the venom gland over our threshold of 100 reads. In snakes, over 80 C-type lectin venom peptides affecting hemostatsis and thrombosis have been identified (Ogawa et al. 2005). Lectins have also been found in the venoms of caterpillars, blood-sucking insects, and stonefish (Fry et al. 2009). Interestingly, the platypus has experienced a massive expansion of C-type lectin domain genes (Wong et al. 2009). Due to their similarity to NKC receptors, it was thought that these genes primarily have immunological function, but evidence of high expression in venom gland warrants further exploration of a venom function.

Evidence of Adaptive Selection

We do not expect that all genes under positive selection will have venom-related roles. Rather, the majority appear to be associated with ubiquitous cellular processes such as transcription and cell metabolism. Among the 20 pairs of gene duplicates that showed evidence of positive selection, two pairs belonged to gene families with a known venom function (table 2): a disintegrin and metalloproteinase (ADAM) family protein (ENSOANP00000007182) and a C-type lectin protein (ENSOANP00000011236). Genes from the ADAM gene family are closely related to snake venom metalloproteinases (SVMPs), which have been shown to evolve through positive selection (Casewell, Wagstaff, Harrison, Renjifo, et al. 2011). SVMPs can be cleaved into disintegrin peptides with potent inhibition of platelet aggregation activity but can also have hemorrhagic activities (Jia et al. 1996; Silva et al. 2003; Modesto et al. 2005). Based on the coagulatory properties of platypus venom, we suggest that the platypus peptide likely acts in a procoagulatory capacity. We did not identify a predicted signal peptide in the platypus ADAM peptide using SignalP. This is not unexpected, as it is thought that certain venom metalloproteinases possess an internal signal peptide in the form of a hydrophobic segment which only acts as a signal peptide in the mature peptide following proteolysis of the precursor (Kini 1995). Platypus C-type lectin genes have previously been shown to evolve under positive selection (Wong et al. 2009). However, this is the first time that positively selected C-type lectins have been identified in the venom gland. In snakes, C-type lectins have also evolved adaptively and have acquired numerous hemostatic and thrombotic functions (Ogawa et al. 2005). We expect these platypus lectins to possess coagulatory and hypotensive roles due to the observed symptoms of envenomation. We also identified that the TRP channel receptor gene, ENSOANP00000019830, possibly involved in venom secretion, is also under adaptive selection.

In total, we found that approximately 12% of all platypus venom-expressed lineage-specific duplicates showed evidence of positive selection. This is likely to be an underestimate for several reasons. There is a limited time period postduplication when positive selection can be detected, and signatures of positive selection can be obscured by purifying selection in older gene duplicates (Han et al. 2009). In addition, positive selection likely acts on a small number of amino acids for a short period of time after duplication and therefore is most detectable in younger duplicates (Zhang 2003). Positive selection is also difficult to detect in recently diverged genes. The chi-square test, used to test the significance of the model of positive selection, also gives conservative results for short and highly similar data sets (Anisimova et al. 2001). Moreover, it is generally difficult to detect positive selection between gene pairs due to the small numbers of sites available for comparison, resulting in little discriminatory power. Despite this, our result is comparable to the amount of positive selection in lineage-specific gene duplicates in four eutherian lineages: human, macaque, rat, and mouse (Han et al. 2009). There, the authors only examined “young” duplicates, those with ds < 1, enriching their data set for genes that show evidence of adaptive selection (Lynch and Conery 2000; Kondrashov et al. 2002). In comparison, approximately 8% of our duplicates possessed a ds value greater than 1.

Limited Role of Gene Duplications

We did not find instances where all gene members of a large multigene family expansion were expressed in the venom gland. We identified expanded lineage-specific clades for ADAM peptides, serine proteases, Kunitz-type serine proteases inhibitors, and C-type lectin genes, but only some and not all members of the expanded clade were expressed in the venom gland. This suggests multiple recruitment events from the same gene family without further duplications that increase toxin copies. Extensive gene duplications can lead to increase in toxin expression, thereby increasing the efficiency and potency of the venom (Fry et al. 2009). Multiple gene copies can increase toxin concentration and hasten venom replacement. Cone snail toxins use this pattern of evolution; where between 50 and 100 thousand individual peptides have been identified across a number of species (Bingham et al. 2010). There may be several reasons why we did not observe expansions of multivenom copies in platypus. Rapid venom replenishment may not be required in the platypus, as only small quantities of venom are released (70 μl) and the full venom gland capacity is not fully delivered by the animal when it spurs (Whittington et al. 2009). Additionally, the role of platypus venom is believed to be for asserting dominance over conspecifics in the context of mating rather than to kill or immobilize prey, as in cone snails and snakes. These functional differences may account for differences in strength and direction of selective pressure on venom genes as it is likely that there is an increased fitness cost in producing lethal venom as opposed to a less potent venom that incapacitates their mating competitors. However, we note that lowly expressed genes, incorrect phylogenetic assignments or inaccurate gene predictions may have affected our ability to detect multiple toxin copies in the platypus.

Most platypus venom genes do not appear to have arisen via gene duplication in the monotreme lineage. Only 16 of 107 putative venom genes identified through similarity to known toxin genes are lineage-specific duplicates. If we assume convergent evolution of venom function, only 15% of platypus toxins have evolved through gene duplications. Consistent with this, nine of 12 serine proteases identified by Whittington et al. (2010) have one-to-one orthologs with other species; in addition, platypus venom protein fractions, C-type natriuretic peptides (ENSOANG00000012322) (de Plater et al. 1998) and nerve growth factor (most likely to be ENSOANG00000004523) (de Plater et al. 1995) identified previously via proteomic studies do not appear to have duplicated in the platypus lineage.

We suggest that changes other than gene duplications play an important part in the recruitment of toxin genes. There are 12 known splice variants of platypus venom C-type natriuretic peptide (Kita et al. 2009). Snake venom C-type natriuretic peptides also undergo extensive molecular processing forming a large number of isoforms from a single precursor mRNA (Higuchi et al. 1999). This extensive cleavage pattern has also been identified in viper sarafotoxins (Ducancel et al. 1993). Alternatively, spliced variants have been identified in other venom components including snake disintegrins (Scarborough et al. 1993) and Factor X (St. Pierre et al. 2005). Splice variants are considered to be a major contributor to protein diversity (Modrek and Lee 2002; Keren et al. 2010). Human genes (92–94%) undergo alternative splicing and are responsible for the 4-fold increase in the number of proteins synthesized compared with protein-coding genes (Wang et al. 2008; Keren et al. 2010). Protein isoforms derived from alternatively spliced genes can have distinct or even antagonistic functions (Das et al. 2002; Zhang and Mount 2009) and may possess different tissue specificities (Wang et al. 2008). Gene mutations can also increase the functional diversity of a gene, through changes in either the regulatory or the protein-coding regions. In snakes, CRISP and kallikrein venom genes are believed to have evolved through mutational modifications (Fry 2005).

It is also possible that we have underestimated the number of putative venom genes that have undergone duplication. This could be due to several factors. First, limited gene divergence or short gene length could lead to difficulties identifying duplicates. For example, previously gene trees were not constructed for defensin-like venom peptides (Whittington et al. 2008) due to their high degree of divergence. Second, although a large number of protein-coding genes were predicted in the platypus assembly (17,951), it is possible that some venom gene duplicates were missed. Third, the fragmented assembly is likely to have led to truncations in some gene predictions and hence failure to call some duplications.

Gene duplications are only partially responsible for the adaptation of platypus venom. The availability of whole genome sequence has allowed us to detect gene duplicates that have led to venom function. Multispecies comparisons have allowed us to distinguish between duplicates that stem from speciation events, where venom orthologs exist in nonvenomous species and true lineage-specific innovations. Comparative studies of other venomous animals will provide insights on the likely variable role of gene duplication in venom evolution. We anticipate that genomic strategies will play an increasingly useful role in the understanding of venom evolution as whole genome sequences from other venomous animals become available.


Our genomic and transcriptomic data analysis shows that duplications of proteases, their inhibitors, and C-type lectins play a role in the evolution of platypus venom. However, extensive venom gene duplications were not identified and key venom genes do not appear to have arisen through gene duplications in the platypus lineage. This points to other evolutionary mechanisms being important in platypus venom evolution. Our study highlights the usefulness of genomic data in understanding the evolution of novel gene functions.

Supplementary Material

Supplementary files 1 and 2 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).


This work was funded by the University of Sydney and the Australian Research Council. KB is supported by an ARC Future Fellowship.


  • Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–3402. [PMC free article] [PubMed]
  • Aminetzach YT, Srouji JR, Kong CY, Hoekstra HE. Convergent evolution of novel protein function in shrew and lizard venom. Curr Biol. 2009;19:1925–1931. [PubMed]
  • Anisimova M, Bielawski JP, Yang Z. Accuracy and power of the likelihood ratio test in detecting adaptive molecular evolution. Mol Biol Evol. 2001;18:1585–1592. [PubMed]
  • Anisimova M, Nielsen R, Yang Z. Effect of recombination on the accuracy of the likelihood method for detecting positive selection at amino acid sites. Genetics. 2003;164:1229–1236. [PMC free article] [PubMed]
  • Ashburner M, Ball CA, Blake JA, et al. (20 co-authors) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000;25:25–29. [PMC free article] [PubMed]
  • Bailey JA, Gu Z, Clark RA, Reinert K, Samonte RV, Schwartz S, Adams MD, Myers EW, Li PW, Eichler EE. Recent segmental duplications in the human genome. Science. 2002;297:1003–1007. [PubMed]
  • Bendtsen JD, Nielsen H, von Heijne G, Brunak S. Improved prediction of signal peptides: SignalP 3.0. J Mol Biol. 2004;340:783–795. [PubMed]
  • Bergthorsson U, Andersson DI, Roth JR. Ohno's dilemma: evolution of new genes under continuous selection. Proc Natl Acad Sci U S A. 2007;104:17004–17009. [PMC free article] [PubMed]
  • Biardi JE, Chien DC, Coss RG. California ground squirrel (Spermophilus beecheyi) defenses against rattlesnake venom digestive and hemostatic toxins. J Chem Ecol. 2006;32:137–154. [PubMed]
  • Bingham J-P, Mitsunaga E, Bergeron ZL. Drugs from slugs—past, present and future perspectives of [omega]-conotoxin research. Chem Biol Interact. 2010;183:1–18. [PubMed]
  • Calvete JJ, Sanz L, Angulo Y, Lomonte B, Gutiérrez JM. Venoms, venomics, antivenomics. FEBS Lett. 2009;583:1736–1743. [PubMed]
  • Casewell NR, Wagstaff SC, Harrison RA, Renjifo C, Wüster W. Domain loss facilitates accelerated evolution and neofunctionalization of duplicate snake venom metalloproteinase toxin genes. Mol Biol Evol. 2011a [cited 2011 April 4]. PMID 21498373. [PubMed]
  • Casewell NR, Wagstaff SC, Harrison RA, Wüster W. Gene tree parsimony of multilocus snake venom protein families reveals species tree conflict as a result of multiple parallel gene loss. Mol Biol Evol. 2011b;28:1157–1172. [PubMed]
  • Clapham DE, Runnels LW, Strübing C. The TRP ion channel family. Nat Rev Neurosci. 2001;2:387–396. [PubMed]
  • Clark DP. Molecular biology. Burlington (MA): Academic Press; 2005.
  • Colinet D, Dubuffet A, Cazes D, Moreau S, Drezen J-M, Poirié M. A serpin from the parasitoid wasp Leptopilina boulardi targets the Drosophila phenoloxidase cascade. Dev Comp Immunol. 2009;33:681–689. [PubMed]
  • Daltry JC, Wuster W, Thorpe RS. Diet and snake venom evolution. Nature. 1996;379:537–540. [PubMed]
  • Das S, Lin J-H, Papamatheakis J, Sykulev Y, Tsichlis PN. Differential splicing generates Tvl-1/RFXANK isoforms with different functions. J Biol Chem. 2002;277:45172–45180. [PubMed]
  • de Plater G, Martin RL, Milburn PJ. A pharmacological and biochemical investigation of the venom from the platypus (Ornithorhynchus anatinus) Toxicon. 1995;33:157–169. [PubMed]
  • de Plater GM, Martin RL, Milburn PJ. A C-type natriuretic peptide from the venom of the platypus (Ornithorhynchus anatinus): structure and pharmacology. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol. 1998;120:99–110. [PubMed]
  • Deshimaru M, Ogawa T, Nakashima K, et al. (11 co-authors) Accelerated evolution of crotalinae snake venom gland serine proteases. FEBS Lett. 1996;397:83–88. [PubMed]
  • Drüeke TB. Extraskeletal problems and amyloid. Kidney Int Suppl. 1999;73:S89–S93. [PubMed]
  • Ducancel F, Matre V, Dupont C, Lajeunesse E, Wollberg Z, Bdolah A, Kochva E, Boulain JC, Ménez A. Cloning and sequence analysis of cDNAs encoding precursors of sarafotoxins. Evidence for an unusual “rosary-type” organization. J Biol Chem. 1993;268:3052–3055. [PubMed]
  • Duda TF, Palumbi SR. Molecular genetics of ecological diversification: duplication and rapid evolution of toxin genes of the venomous gastropod Conus. Proc Natl Acad Sci U S A. 1999;96:6820–6823. [PMC free article] [PubMed]
  • Fenner PJ, Williamson JA, Myers D. Platypus envenomation—a painful learning experience. Med J Aust. 1992;157:829–832. [PubMed]
  • Fletcher W, Yang Z. The effect of insertions, deletions, and alignment errors on the branch-site test of positive selection. Mol Biol Evol. 2010;27:2257–2267. [PubMed]
  • Flicek P, Amode MR, Barrell D, et al. (52 co-authors) Ensembl 2011. Nucleic Acids Res. 2010;39:D800–D806. [PMC free article] [PubMed]
  • Fry BG. From genome to “venome”: molecular origin and evolution of the snake venom proteome inferred from phylogenetic analysis of toxin sequences and related body proteins. Genome Res. 2005;15:403–420. [PMC free article] [PubMed]
  • Fry BG, Roelants K, Champagne DE, et al. (12 co-authors) The toxicogenomic multiverse: convergent recruitment of proteins into animal venoms. Annu Rev Genomics Hum Genet. 2009;10:483–511. [PubMed]
  • Fry BG, Wuster W, Kini RM, Brusic V, Khan A, Venkataraman D, Rooney AP. Molecular evolution and phylogeny of elapid snake venom three-finger toxins. J Mol Evol. 2003;57:110–129. [PubMed]
  • Grant T. Platypus. 4th ed. Collingwood: CSIRO Publishing; 2007.
  • Gutiérrez J, Lomonte B. Phospholipase A2 myotoxins from Bothrops snake venoms. Toxicon. 1995;33:1405–1424. [PubMed]
  • Han MV, Demuth JP, McGrath CL, Casola C, Hahn MW. Adaptive evolution of young gene duplicates in mammals. Genome Res. 2009;19:859–867. [PMC free article] [PubMed]
  • Hattori A, Kitatani K, Matsumoto H, Miyazawa S, Rogi T, Tsuruoka N, Mizutani S, Natori Y, Tsujimoto M. Characterization of recombinant human adipocyte-derived leucine aminopeptidase expressed in Chinese hamster ovary cells. J Biochem. 2000;128:755–762. [PubMed]
  • Higuchi S, Murayama N, Saguchi K, Ohi H, Fujita Y, Camargo ACM, Ogawa T, Deshimaru M, Ohno M. Bradykinin-potentiating peptides and C-type natriuretic peptides from snake venom. Immunopharmacology. 1999;44:129–135. [PubMed]
  • Hirakura Y, Kagan BL. Pore formation by beta-2-microglobulin: a mechanism for the pathogenesis of dialysis associated amyloidosis. Amyloid. 2001;8:94–100. [PubMed]
  • Hochberg Y, Benjamini Y. More powerful procedures for multiple significance testing. Stat Med. 1990;9:811–818. [PubMed]
  • Jia L-G, Shimokawa K-I, Bjarnason JB, Fox JW. Snake venom metalloproteinases: structure, function and relationship to the ADAMs family of proteins. Toxicon. 1996;34:1269–1276. [PubMed]
  • Juárez P, Comas I, González-Candelas F, Calvete JJ. Evolution of snake venom disintegrins by positive Darwinian selection. Mol Biol Evol. 2008;25:2391–2407. [PubMed]
  • Jungo F, Bairoch A. Tox-Prot, the toxin protein annotation program of the Swiss-Prot protein knowledgebase. Toxicon. 2005;45:293–301. [PubMed]
  • Keren H, Lev-Maor G, Ast G. Alternative splicing and evolution: diversification, exon definition and function. Nat Rev Genet. 2010;11:345–355. [PubMed]
  • Kini RM. Do we know the complete sequence of metalloproteinase and nonenzymatic platelet aggregation inhibitor (disintegrin) precursor proteins? Toxicon. 1995;33:1151–1160. [PubMed]
  • Kini RM. Serine proteases affecting blood coagulation and fibrinolysis from snake venoms. Pathophysiol Haemost Thromb. 2005;34:200–204. [PubMed]
  • Kita M, Black DS, Ohno O, Yamada K, Kigoshi H, Uemura D. Duck-billed platypus venom peptides induce Ca2+ influx in neuroblastoma cells. J Am Chem Soc. 2009;131:18038–18039. [PubMed]
  • Kita M, Nakamura Y, Okumura Y, Ohdachi SD, Oba Y, Yoshikuni M, Kido H, Uemura D. Blarina toxin, a mammalian lethal venom from the short-tailed shrew Blarina brevicauda: isolation and characterization. Proc Natl Acad Sci U S A. 2004;101:7542–7547. [PMC free article] [PubMed]
  • Kondrashov FA, Rogozin IB, Wolf YI, Koonin EV. Selection in the evolution of gene duplications. Genome Biol. 2002;3:RESEARCH0008. [PMC free article] [PubMed]
  • Kordis D, Gubensek F. Adaptive evolution of animal toxin multigene families. Gene. 2000;261:43–52. [PubMed]
  • Kress LF, Catanese JJ. Effects of human plasma [alpha]2 macroglobulin on the proteolytic activity of snake venoms. Toxicon. 1981;19:501–507. [PubMed]
  • Li H. Maq: Mapping and Assembly with Qualities. Available from: http://maq.sourceforge.net/faq.shtml.
  • Löytynoja A, Goldman N. Phylogeny-aware gap placement prevents errors in sequence alignment and evolutionary analysis. Science. 2008;320:1632–1635. [PubMed]
  • Luban J, Bossolt KL, Franke EK, Kalpana GV, Goff SP. Human immunodeficiency virus type 1 Gag protein binds to cyclophilins A and B. Cell. 1993;73:1067–1078. [PubMed]
  • Lynch M, Conery JS. The evolutionary fate and consequences of duplicate genes. Science. 2000;290:1151–1155. [PubMed]
  • Lynch VJ. Inventing an arsenal: adaptive evolution and neofunctionalization of snake venom phospholipase A2 genes. BMC Evol Biol. 2007;7:2. [PMC free article] [PubMed]
  • Mamathambika BS, Bardwell JC. Disulfide-linked protein folding pathways. Annu Rev Cell Dev Biol. 2008;24:211–235. [PubMed]
  • Martin CJ, Tidswell F. Observations on the femoral gland of Ornithorhynchus and its secretion; together with an experimental enquiry concerning its supposed toxic action. Proc Linn Soc N S W. 1895;9:471–500.
  • Minh Le TN, Reza MA, Swarup S, Kini RM. Gene duplication of coagulation factor V and origin of venom prothrombin activator in Pseudonaja textilis snake. Thromb Haemost. 2005;93:420–429. [PubMed]
  • Modesto JC, Junqueira-de-Azevedo IL, Neves-Ferreira AG, Fritzen M, Oliva ML, Ho PL, Perales J, Chudzinski-Tavassi AM. Insularinase A, a prothrombin activator from Bothrops insularis venom, is a metalloprotease derived from a gene encoding protease and disintegrin domains. Biol Chem. 2005;386:589–600. [PubMed]
  • Modrek B, Lee C. A genomic view of alternative splicing. Nat Genet. 2002;30:13–19. [PubMed]
  • Murray SA, Mihali TK, Neilan BA. Extraordinary conservation, gene loss, and positive selection in the evolution of an ancient neurotoxin. Mol Biol Evol. 2011;28:1173–1182. [PubMed]
  • Ogawa T, Chijiwa T, Oda-Ueda N, Ohno M. Molecular diversity and accelerated evolution of C-type lectin-like proteins from snake venom. Toxicon. 2005;45:1–14. [PubMed]
  • Ohno S. Evolution by gene duplication. New York: Springer; 1970.
  • Poran NS, Coss RG, Benjamini E. Resistance of California ground squirrels (Spermophilus beecheyi) to the venom of the northern Pacific rattlesnake (Crotalus viridis oreganus): a study of adaptive variation. Toxicon. 1987;25:767–777. [PubMed]
  • Reza MA, Minh Le TN, Swarup S, Manjunatha Kini R. Molecular evolution caught in action: gene duplication and evolution of molecular isoforms of prothrombin activators in Pseudonaja textilis (brown snake) J Thromb Haemost. 2006;4:1346–1353. [PubMed]
  • Roelants K, Fry BG, Norman JA, Clynen E, Schoofs L, Bossuyt F. Identical skin toxins by convergent molecular adaptation in frogs. Curr Biol. 2010;20:125–130. [PubMed]
  • Rubin H, Wang ZM, Nickbarg EB, McLarney S, Naidoo N, Schoenberger OL, Johnson JL, Cooperman BS. Cloning, expression, purification, and biological activity of recombinant native and variant human alpha 1-antichymotrypsins. J Biol Chem. 1990;265:1199–1207. [PubMed]
  • Rycyzyn MA, Clevenger CV. The intranuclear prolactin/cyclophilin B complex as a transcriptional inducer. Proc Natl Acad Sci U S A. 2002;99:6790–6795. [PMC free article] [PubMed]
  • Safavi-Hemami H, Bulaj G, Olivera BM, Williamson NA, Purcell AW. Identification of Conus peptidylprolyl cis-trans isomerases (PPIases) and assessment of their role in the oxidative folding of conotoxins. J Biol Chem. 2010;285:12735–12746. [PMC free article] [PubMed]
  • Sawyer SA. GENECONV: a computer package for the statistical detection of gene conversion. Washington University in St. Louis; 1999. Available from: http://www.math.wustl.edu/~sawyer.
  • Scarborough RM, Rose JW, Naughton MA, Phillips DR, Nannizzi L, Arfsten A, Campbell AM, Charo IF. Characterization of the integrin specificities of disintegrins isolated from American pit viper venoms. J Biol Chem. 1993;268:1058–1065. [PubMed]
  • Silva MB, Schattner M, Ramos CRR, et al. (11 co-authors) A prothrombin activator from Bothrops erythromelas (jararaca-da-seca) snake venom: characterization and molecular cloning. Biochem J. 2003;369:129–139. [PMC free article] [PubMed]
  • Sottrup-Jensen L, Stepanik TM, Kristensen T, et al. (12 co-authors) Common evolutionary origin of α 2-macroglobulin and complement components C3 and C4. Proc Natl Acad Sci U S A. 1985;82:9–13. [PMC free article] [PubMed]
  • St. Pierre L, Masci PP, Filippovich I, Sorokina N, Marsh N, Miller DJ, Lavin MF. Comparative analysis of prothrombin activators from the venom of Australian elapids. Mol Biol Evol. 2005;22:1853–1864. [PubMed]
  • Tamura K, Dudley J, Nei M, Kumar S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol. 2007;24:1596–1599. [PubMed]
  • Torres AM, Tsampazi M, Kennett EC, Belov K, Geraghty DP, Bansal PS, Alewood PF, Kuchel PW. Characterization and isolation of L-to-D-amino-acid-residue isomerase from platypus venom. Amino Acids. 2007;32:63–68. [PubMed]
  • Torres AM, Tsampazi M, Tsampazi C, Kennett EC, Belov K, Geraghty DP, Bansal PS, Alewood PF, Kuchel PW. Mammalian l-to-d-amino-acid-residue isomerase from platypus venom. FEBS Lett. 2006;580:1587–1591. [PubMed]
  • van Gent D, Sharp P, Morgan K, Kalsheker N. Serpins: structure, function and molecular evolution. Int J Biochem Cell Biol. 2003;35:1536–1547. [PubMed]
  • Vilella AJ, Severin J, Ureta-Vidal A, Heng L, Durbin R, Birney E. EnsemblCompara GeneTrees: complete, duplication-aware phylogenetic trees in vertebrates. Genome Res. 2009;19:327–335. [PMC free article] [PubMed]
  • Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, Kingsmore SF, Schroth GP, Burge CB. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008;456:470–476. [PMC free article] [PubMed]
  • Warren WC, Hillier LW, Marshall Graves JA, et al. (104 co-authors) Genome analysis of the platypus reveals unique signatures of evolution. Nature. 2008;453:175–183. [PMC free article] [PubMed]
  • Whittington CM, Koh JMS, Warren WC, Papenfuss AT, Torres AM, Kuchel PW, Belov K. Understanding and utilising mammalian venom via a platypus venom transcriptome. J Proteomics. 2009;72:155–164. [PubMed]
  • Whittington CM, Papenfuss AT, Bansal P, et al. (15 co-authors) Defensins and the convergent evolution of platypus and reptile venom genes. Genome Res. 2008;18:986–994. [PMC free article] [PubMed]
  • Whittington CM, Papenfuss AT, Locke DP, et al. (12 co-authors) Novel venom gene discovery in the platypus. Genome Biol. 2010;11:R95. [PMC free article] [PubMed]
  • Wong ESW, Sanderson CE, Deakin JE, Whittington CM, Papenfuss AT, Belov K. Identification of natural killer cell receptor clusters in the platypus genome reveals an expansion of C-type lectin genes. Immunogenetics. 2009;61:565–579. [PubMed]
  • Wu Q, Kuo H-C, Deng GG. Serine proteases and cardiac function. Biochim Biophys Acta. 2005;1751:82–94. [PubMed]
  • Yang Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Mol Biol Evol. 2007;24:1586–1591. [PubMed]
  • Zhang J. Evolution by gene duplication: an update. Trends Ecol Evol. 2003;18:292–298.
  • Zhang X-N, Mount SM. Two alternatively spliced isoforms of the Arabidopsis SR45 protein have distinct roles during normal plant development. Plant Physiol. 2009;150:1450–1458. [PMC free article] [PubMed]
  • Zhijian C, Feng L, Yingliang W, Xin M, Wenxin L. Genetic mechanisms of scorpion venom peptide diversification. Toxicon. 2006;47:348–355. [PubMed]
  • Zupunski V, Kordis D, Gubensek F. Adaptive evolution in the snake venom Kunitz/BPTI protein family. FEBS Lett. 2003;547:131–136. [PubMed]

Articles from Molecular Biology and Evolution are provided here courtesy of Oxford University Press
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...