Abstract

Long-term specialization may limit the ability of a species to respond to new environmental conditions and lead to a higher likelihood of extinction. For permanent parasites and other symbionts, the most intriguing question is whether these organisms can return to a free-living lifestyle and, thus, escape an evolutionary “dead end.” This question is directly related to Dollo's law, which stipulates that a complex trait (such as being free living vs. parasitic) cannot re-evolve again in the same form. Here, we present conclusive evidence that house dust mites, a group of medically important free-living organisms, evolved from permanent parasites of warm-blooded vertebrates. A robust, multigene topology (315 taxa, 8942 nt), ancestral character state reconstruction, and a test for irreversible evolution (Dollo's law) demonstrate that house dust mites have abandoned a parasitic lifestyle, secondarily becoming free living, and then speciated in several habitats. Hence, as exemplified by this model system, highly specialized permanent parasites may drastically de-specialize to the extent of becoming free living and, thus escape from dead-end evolution. Our phylogenetic and historical ecological framework explains the limited cross-reactivity between allergens from the house dust mites and “storage” mites and the ability of the dust mites to inhibit host immune responses. It also provides insights into how ancestral features related to parasitism (frequent ancestral shifts to unrelated hosts, tolerance to lower humidity, and pre-existing enzymes targeting skin and keratinous materials) played a major role in reversal to the free-living state. We propose that parasitic ancestors of pyroglyphids shifted to nests of vertebrates. Later the nest-inhabiting pyroglyphids expanded into human dwellings to become a major source of allergens. [Ancestral ecology; Dollo's law; evolutionary “dead end”; house dust mites; permanent parasitism; Pyroglyphidae.]

“Once a parasite, always a parasite.” This seeming truism appears in recent online discussions of everything from noxious computer software, to welfare cheats, to politicians. Curiously, in evolutionary biology, there is also a strongly rooted supposition that highly specialized traits or ecologies, such as permanent (or full-time) parasitism, result in irreversible or unidirectional evolution (Futuyma and Moreno 1988; Agnarsson et al. 2006; Cruickshank and Paterson 2006; Goldberg and Igic 2008). Although the term “parasite” has many definitions, all involve some degree of dependence between organisms and lasting exploitation of one organism by another, with permanent parasites spending their entire lives on or in the body of a host (Price 1980). The natural outcome of this situation is that parasites may quickly evolve highly sophisticated mechanisms for host exploitation and lose their ability to function away from the host body. Parasites often experience seemingly irreplaceable degradation or loss of many genes as their functionality is no longer required in a rich, predictable environment where hosts provide both the living space and nutrition (Sakharkar et al. 2004; Dieterich and Sommer 2009; Mendonca et al. 2011). Many perceive such specialization as evolutionarily irreversible and, consequently, believe that parasites can never return to their ancestrally free-living lifestyle again (Toft et al. 1991; Combes and Simberloff 2005).

High degrees of specialization may evolve by short-term selective advantage. However, long-term specialization may limit the ability to respond to new environmental conditions leading to a higher likelihood of extinction and the proverbial evolutionary “dead end” (Takebayashi and Morrell 2001; Termonia et al. 2001; Agnarsson et al. 2006). The question of whether or not highly specialized lineages can enter a different ecological niche, and thus, avoid the fate of evolutionary “dead ends” is a subject of ongoing debates (reviewed in Colles et al. 2009). Several studies suggest that various organisms indeed can offset the constraints imposed by specialization (Holmes 1977; Lanyon 1992; Armbruster and Baldwin 1998; D'Haese 2000; Janz et al. 2001; Termonia et al. 2001; Morse and Farrell 2005; Nosil and Mooers 2005; Stireman 2005; Colles et al. 2009; Prendini et al. 2010). Particularly, supposed irreversibility of parasitism has been placed under severe scrutiny, resulting in a number of empirical studies that have seemingly brought this parasitological dogma into disrepute (Siddall et al. 1993; Radovsky et al. 1997; Smith 1998; Apakupakul et al. 1999; Light and Siddall 1999; Dorris et al. 2002; Salewski 2003; Borda and Siddall 2004; Cruickshank and Paterson 2006). However, advances in analytical approaches suggest that simple ancestral character state inference, as done in these studies, may be confounded if irreversible evolution (Dollo's law) is assumed (Stireman 2005; Goldberg and Igic 2008).

Standard Markov (mk) models of trait evolution might wrongly reject the hypothesis of irreversible evolution if the net diversification rate (speciation minus extinction) is greater for lineages having the ancestral state of a character subject to Dollo's law (Goldberg and Igic 2008). In cases of parasitism, it means that if free-living lineages (ancestral state) diversify faster than parasitic ones, then on the phylogeny, a small portion of them may appear surrounded by numerous parasitic lineages. In this situation, “standard” methods of ancestral character state reconstruction will wrongly infer reversal of an ancestral state. Therefore, methods accounting for speciation, extinction, and character state transition rates should be employed to test for irreversible evolution (Goldberg and Igic 2008). In this analytical framework, one influential study rejecting Dollo's law under the mk model was demonstrated to be inconclusive (Goldberg and Igic 2008), but a few violations of Dollo's law have since been reported (Kohlsdorf et al. 2010; Lynch and Wagner 2010; Wiens 2011). The question of whether a free-living state can re-evolve in permanent parasites is still in doubt, and indisputable examples are currently lacking.

Here, we analyze the long-standing problem of the ancestral ecology of house dust mites (family Pyroglyphidae), a medically important, yet understudied group of microscopic arthropods. Pyroglyphid house dust mites are the most common cause of allergic symptoms in humans, affecting 65 million to 1.2 billion people worldwide (Cunnington and Gregory 1968; Basagana et al. 2004; Holt and Thomas 2005; Colloff 2009; Hammad et al. 2009; Lloyd 2009). Dust mites feed on organic debris (such as shed skin) and flourish in nests of vertebrates, including human dwellings. The mites' severe allergenic properties are linked to their powerful digestive and molting enzymes, salivary secretions, and other water-soluble molecules. For example, large quantities of digestive enzymes accumulate in mites' fecal material—minute particles that can easily become airborne. When inhaled or in contact with skin, the residual mite enzymes may break up tight junctions between the epithelial cells and trigger allergic reactions (Tovey et al. 1981; Arlian 1991; Holt and Thomas 2005; Hammad et al. 2009; Lloyd 2009).

Pyroglyphid mites belong to a large acarine lineage, the Astigmata, comprising more than 6100 described species. This lineage contains an ensemble of several mutually paraphyletic, mostly free-living lineages (“free-living Astigmata”) and a large monophyletic parasitic lineage—Psoroptidia.

Free-living Astigmata (ca. 2300 species) are usually saprophagous, feeding on decaying organic material, fungi, and bacteria found in patchy or ephemeral habitats—fungal sporocarps, sap flows, dung, carrion, seaweeds, insect and vertebrate nests, and stored food (hence the colloquial name “storage mites”). Storage mites often develop large populations in dust and stored biological products (e.g., grain, straw), and can cause allergies in humans (Sidenius et al. 2001; Ong and Chew 2010). Specialized deutonymphs often disperse among habitat patches through phoretic associations with arthropods and vertebrates that also use these resources. Phoretic deutonymphs usually do not feed during dispersal, but there are several unrelated free-living lineages with parasitic deutonymphs: Hypoderatidae, Glycyphagidae, Chortoglyphidae, and Echimyopodidae (Fig. 1, color-coded yellow). These deutonymphs occur inside hair follicles or under the skin, where they acquire nutrition from hosts (the exact mechanism is still not understood—these deutonymphs do not have a mouth or normal gut).

Figure 1

Phylogram of pyroglyphid house dust mites and outgroups inferred by ML analysis a) (315 taxa, 6164 sites, nuclear rDNA and protein-coding genes translated into amino acids). Four numbered key nodes leading to the dust mites are supplemented with bootstrap support values, posterior probabilities (BA), and pie charts showing results of ancestral character state reconstruction for ecology (parasitic and parasitic as deutonymphs vs. free-living) for these nodes (here values for Bayesian 1-rate Markov model are shown, see Table 2 for probabilities calculated using other methods). Node 1 = Hypoderatidae+Heterocoptidae+Psoroptidia; 2 = Heterocoptidae+Psoroptidia; 3 = Psoroptidia; 4 = EP complex. b) ML tree of HSP70 (315 taxa, 569 amino acids, part of tree shown). c) BEAST species tree analysis (46 taxa, 4146 nt, protein-coding genes only, part of tree shown). The inferences are uncertain if in the dust mites a free-living state evolved twice a) or once as suggested by the HSP70 topology b) or species tree topology inferred from protein-coding genes c). Both scenarios have low support. Scale bars represent expected changes per site. Bootstrap support or posterior probability of 70–100% is shown by proportionally increased line weight (see actual values and taxon names in Supplementary Figs. S3–S5).

Figure 1

Phylogram of pyroglyphid house dust mites and outgroups inferred by ML analysis a) (315 taxa, 6164 sites, nuclear rDNA and protein-coding genes translated into amino acids). Four numbered key nodes leading to the dust mites are supplemented with bootstrap support values, posterior probabilities (BA), and pie charts showing results of ancestral character state reconstruction for ecology (parasitic and parasitic as deutonymphs vs. free-living) for these nodes (here values for Bayesian 1-rate Markov model are shown, see Table 2 for probabilities calculated using other methods). Node 1 = Hypoderatidae+Heterocoptidae+Psoroptidia; 2 = Heterocoptidae+Psoroptidia; 3 = Psoroptidia; 4 = EP complex. b) ML tree of HSP70 (315 taxa, 569 amino acids, part of tree shown). c) BEAST species tree analysis (46 taxa, 4146 nt, protein-coding genes only, part of tree shown). The inferences are uncertain if in the dust mites a free-living state evolved twice a) or once as suggested by the HSP70 topology b) or species tree topology inferred from protein-coding genes c). Both scenarios have low support. Scale bars represent expected changes per site. Bootstrap support or posterior probability of 70–100% is shown by proportionally increased line weight (see actual values and taxon names in Supplementary Figs. S3–S5).

Psoroptidia (ca. 3800 species) are permanent or full-time parasites. This group is represented by highly specialized parasites that spend their entire lives on or in a suitable vertebrate host, unable to complete their life cycle without it, and lack a free-living dispersal stage. Host specificity is relatively high, with 80% of mite species specific to host species or genera. The typical host range includes birds and mammals; occasionally they are hyperparasitic on insect parasites. Psoroptidian mites include several ecological groups (respiratory endoparasites, quill, skin-surface and skin-burrowing mites, Fig. 2c–i) that are conventional parasites causing negative fitness effects on their hosts. However, some mites feeding on uropygial gland secretions (serving to waterproof and protect bird plumage) may inflict minimal or negligible harm to the host fitness. They not only rarely cause pathology (reviewed by Proctor 2003) but also can opportunistically feed on blood (PBK personal observation). Sometimes called commensals, such feather vane-dwelling mites (Fig. 2a,b) fit within the classical definition of parasitism because they consume a functionally important component of the body and are unable to survive away from their hosts. However, if these mites are not accepted as being true parasites, then it would be more appropriate to define our research question as “Is full-time symbiosis evolutionarily reversible?”

Relationships of the pyroglyphid dust mites and the nonpsoroptidian and psoroptidian Astigmata are poorly understood. Sixty-two different published hypotheses argue whether pyroglyphids originated from (1) a free-living ancestor (e.g., storage mites) that also gave rise to parasitic lineages or (2) a parasitic ancestor that became secondarily free-living. Although these hypotheses have been discussed on numerous occasions in relation to morphological, molecular, or immunological data (Sidenius et al. 2001; Loo et al. 2003; Suarez-Martinez et al. 2005; Lee et al. 2006; Cui et al. 2008; Klimov and OConnor 2008; Dabert et al. 2010; Ong and Chew 2010; Yang et al. 2010; Bochkov and Mironov 2011), there is no consensus, and none of them have been evaluated statistically. The latter hypothesis represents the most intriguing question—can permanent parasites return to a free-living state over evolutionary time and, thus, escape the fate of being evolutionary “dead ends” in violation of Dollo's law? The most recent treatment dubbed this hypothesis as “impossible” (Bochkov and Mironov 2011).

Figure 2

Specialization to parasitic life style in Psoroptidia. a) Analges sp.: anterior legs with apophyses, spines, and modified setae (arrows) aiding in maneuvering among barbs of undertail covert feathers (museum voucher BMOC 05-1023-001), b) Amerodectes sp.: cylindrical body shape matching the mite restricted environment—spaces between barbs of primary feathers of the wing, the ambulacra (arrows) are enlarged to hold on during the flight (BMOC 84-1002-001), c) Yunkeracarus otomys: extreme reduction of body setae, intranasal parasite of rodents (BMOC 91-1350-114), d) Labidocarpellus eonycteris: legs and gnathosoma are modified (arrows) to attach to the host hair, associated with bats (MAH 85-0131-001), e) Listrophorus squamiferus: body elongated, legs and gnathosoma are modified (arrows) to attach to the host hair, associated with rodents (BMOC 93-1010-001), f) Myocoptes japonensis: legs III (arrow) are modified to attach to the host hair, skin parasite of rodents (BMOC 93-1010-001), g) Knemidokoptes jamaicensis: legs are extremely short, body spherical, skin-burrowing on birds' legs (BMOC 99-0420-011), h) Sarcoptes scabiei: same as above, ambulacra extremely elongated (arrows), skin-burrowing on humans and domestic animals (BMOC 82-0521-030) and i) Heteropsorus sp.: ambulacra hypertrophied to attach to the upper skin of birds' wings (BMOC 07-1015-093).

Figure 2

Specialization to parasitic life style in Psoroptidia. a) Analges sp.: anterior legs with apophyses, spines, and modified setae (arrows) aiding in maneuvering among barbs of undertail covert feathers (museum voucher BMOC 05-1023-001), b) Amerodectes sp.: cylindrical body shape matching the mite restricted environment—spaces between barbs of primary feathers of the wing, the ambulacra (arrows) are enlarged to hold on during the flight (BMOC 84-1002-001), c) Yunkeracarus otomys: extreme reduction of body setae, intranasal parasite of rodents (BMOC 91-1350-114), d) Labidocarpellus eonycteris: legs and gnathosoma are modified (arrows) to attach to the host hair, associated with bats (MAH 85-0131-001), e) Listrophorus squamiferus: body elongated, legs and gnathosoma are modified (arrows) to attach to the host hair, associated with rodents (BMOC 93-1010-001), f) Myocoptes japonensis: legs III (arrow) are modified to attach to the host hair, skin parasite of rodents (BMOC 93-1010-001), g) Knemidokoptes jamaicensis: legs are extremely short, body spherical, skin-burrowing on birds' legs (BMOC 99-0420-011), h) Sarcoptes scabiei: same as above, ambulacra extremely elongated (arrows), skin-burrowing on humans and domestic animals (BMOC 82-0521-030) and i) Heteropsorus sp.: ambulacra hypertrophied to attach to the upper skin of birds' wings (BMOC 07-1015-093).

We sequenced 5 nuclear genes (8942 nt, after removing introns and unalignable regions) of 315 terminal taxa, including all outgroups ever proposed for house dust mites. Using a robust phylogeny, we (i) evaluated the 62 hypotheses of alternative placement of house dust mites using bootstrap replicates generated from site-wise log-likelihoods calculated for each hypothesis in Consel 1.20 (Shimodaira 2002); (ii) conducted an ancestral character state reconstruction to determine the ancestral ecology of house dust mites in BayesTraits (Pagel et al. 2004) using, separately, maximum likelihood (ML) parametric bootstrap trees and Bayesian stationary trees to account for phylogenetic uncertainty; and (iii) tested for irreversible evolution of parasitism in a recently developed framework (Goldberg and Igic 2008) comparing unconstrained models with those constrained for irreversible evolution, while accounting for speciation, extinction, and character state transition rates in DiversiTree (FitzJohn 2010). These tests provide decisive evidence for the historical ecology of house dust mites and answer the question of whether permanent parasitism is evolutionarily reversible.

Materials and Methods

Taxon Sampling, DNA Isolation, and Sequencing

Five genes, 18S, 28S rDNA, EF1-α, SRP54, and HSP70 (8942 nt aligned, including 4775 nt of rDNA, 1389 amino acids of protein coding genes, no missing data due to sequencing failures), were sequenced for 315 taxa (Supplementary Table S1, available from http://datadryad.org, doi:10.5061/dryad.rd1bc). The total number of nucleotides in our alignment is comparable (55–119%) to recent multigene phyloge-nomic studies (Dunn et al. 2008; Kocot et al. 2011; Struck et al. 2011). As such, our sampling strategy represents the broadly sampled, multigene approach, which has been proven to be very successful in several studies (Parfrey et al. 2010; Rothfels et al. 2012). A total of 1535 new sequences were deposited tin GenBank under accession numbers JQ000032–JQ001566. Another 40 sequences were utilized from our previous studies (Klimov and OConnor 2008; Knowles and Klimov 2011; see Supplementary Table S1). Taxa were selected to account for all previously proposed molecular and morphological hypotheses on dust mite origins (reviewed Klimov and OConnor 2008). The distant outgroups in the acarine order Acariformes were “Endeostigmata” (Lee et al. 2006; Domes et al. 2007) and Oribatida (Norton 1998; Dabert et al. 2010; Pepato et al. 2010). Potential close outgroups were representative sets of nonpsoroptidian (including storage mites) and psoroptidian (associated with birds and mammals) astigmatid mites. Our data set represents nearly 4.8% of known diversity of Astigmata, but it is well balanced, and does not contain obvious taxonomic biases. For example, the percentages of species in the 3 focal groups (free-living Astigmata, nonpyroglyphid Psoroptidia, and pyroglyphids) are comparable with the observed values (26.3 vs. 37.3, 70.0 vs. 61.8, and 3.7 vs. 0.9, respectively). It is almost impossible to obtain a complete taxonomic sampling of astigmatid mites given the large size of this group (6150 known species, representing, perhaps, only 5–10% of the real diversity). The ingroup pyroglyphid taxon sampling was nearly exhaustive at the family–subfamily level and included all common species causing allergies. DNA extraction, rDNA secondary structure alignment, oligonucleotide primers, amplification, and sequencing were previously described (Knowles and Klimov 2011).

Phylogenetic Analyses

Alignment of rDNA was based on secondary structures of Apis mellifera (Gillespie et al. 2006) and Saccharomyces cerevisiae available from the Comparative RNA Web site (Cannone et al. 2002). Stem regions were further evaluated in the program mfold v. 3.1, which folds rRNA based on free energy minimization (Mathews et al. 1999; Zuker 2003) using the default settings. Compensatory mutations in stem regions were detected using a large comparative data set (543 sequences of mites, unpublished). Unalignable, hypervariable regions of rDNA (no common secondary structure can be found) and introns were excluded to avoid erroneous homology assignment. Exons of protein-coding genes, with a few 3- or 6-nt indels, were aligned in MacClade and then checked by eye. Orthology of the genes was unambiguous based on the absence of heterogeneous amplicons. All sequences were checked against the nucleotide database using BLAST (Altschul et al. 1997) to identify potential contaminants. Sequences of protein-coding genes were translated into protein prior to analyses.

Models of nucleotide or amino acid evolution were selected based on corrected Akaike information criterion (AICc) in the programs jModelTest 2.0 (Darriba et al. 2011a) and Prottest 3 (Darriba et al. 2011b), respectively. We explored several partition strategies (by rDNA and amino acids, individual genes, rDNA stem and loop regions) and selected the following partition strategy based on the lowest AICc value: rDNA stem, rDNA loop, EF1-α, SRP54, and HSP70 (Supplementary Table S2). Phylogenetic relationships were inferred in a ML and Bayesian framework in Garli 2.0.1019 (Zwickl 2006) and MrBayes 3.1.2 (Ronquist and Huelsenbeck 2003; Altekar et al. 2004) using a 52-node Mac OS X computer cluster. At least 5 (Garli) or 6 (MrBayes) independent runs were performed. Concatenated and separate analyses for each partition were run in Garli (Supplementary Fig. S3) to estimate potential biases introduced by discordant gene genealogies on the total evidence tree (Degnan and Rosenberg 2009). We used stem and loop regions of rDNA as separate partitions because they have very different levels of saturation (Klimov and OConnor 2008) and, therefore, may provide insights on phylogenetic signal present in the data set.

For Bayesian analyses (BA), convergence of model parameters and topology were assessed by the standard MrBayes convergence diagnostics (i.e., the average standard deviation of split frequencies values below 0.01 and potential scale reduction factor values approaching 1.00) and the program Are We There Yet? (AWTY) (Nylander et al. 2008). Adequacy of the posterior sample size was evaluated through autocorrelation statistics as implemented in Tracer v. 1.5 (Rambaut and Drummond 2009)—all effective sample size values substantially exceeded 200. We ran MrBayes analyses for 10 million generations discarding the first 100 000 trees as burn-in as determined in Tracer. To investigate whether our partitioned BA may be trapped in regions of parameter space characterized by unrealistically long trees (Brown et al. 2010; Marshall 2010), we compared the average post burn-in tree lengths reported by Tracer and the ML tree length estimate. The latter was shorter (9.136 vs. 10.357) and outside of the 95% Bayesian credibility interval (9.967–10.719), suggesting that our Bayesian inference may somehow overestimate branch lengths. However, these estimates may not be completely unrealistic in contrast to studies reporting them as several orders of magnitude longer than corresponding ML estimates, while recovering the same topology (Brown et al. 2010). Given a potential problem with Bayesian branch length estimates, we restrict our study to tests relying on ML topologies. In the only test utilizing Bayesian trees (BayesTraits), results were remarkably similar to those obtained for ML trees.

ML trees were annotated with nonparametric bootstrap support values (104–108 replicates) using the program SumTrees (Sukumaran and Holder 2010) and visualized in FigTree 1.3.1 (Drummond and Rambaut 2007). Matrices and trees from this study are available from TreeBASE (http://www.treebase.org) accession number 12087.

Because monophyly of Dermatophagoidinae+ Pyroglyphinae (major free-living lineages of the dust mites) was not recovered, and because the status of this group has potential implications for the evolutionary loss of parasitism, we investigated these relationships more closely by estimating a species tree using the coalescent-based species tree method in BEAST v1.7.3 (Drummond and Rambaut 2007; Heled and Drummond 2010). For this analysis, a subset of taxa (n = 46), including the epidermoptid–psoroptid (EP) complex (Fig. 1a: node 4) and 8 outgroups of the family Analgidae, was used. A series of analyses utilizing different partitions and their combinations (see above, partitioning strategy by codon position was also used for protein-coding genes) was conducted using at least 2 independent Markov chain Monte Carlo (MCMC) analyses run for 1.2–3.8 ×108 generations with parameters sampled every 1000 steps. Independent runs were combined using the program LogCombiner v. 1.4.6 (Drummond and Rambaut 2007) and burn-in samples were discarded. Convergence and adequacy of the posterior sample size were determined as above for MrBayes analyses.

Hypothesis Testing

Evaluation of 62 previously proposed hypotheses of alternative placement of house dust mites was done in Consel 1.20 using the AU statistic as the primary test (Shimodaira 2002). Consel calculates P-values for the AU and other types of statistics using bootstrap replicates generated from site-wise log-likelihoods with the RELL resampling method—a fast and accurate approximation technique (Shimodaira 2001). We generated a matrix of site-wise log-likelihoods in Garli using constraints corresponding to each of the 62 hypotheses on phylogenetic placement of pyroglyphids. We also report the proportion of trees sampled from the posterior in our preferred BA.

Ancestral character state reconstruction was done in ML and Bayesian frameworks in BayesTraits (Pagel et al. 2004). To account for uncertainty in our phylogenetic inferences, bootstrap trees (ML, 108 nonparametric bootstrap replicates) and post burn-in trees sampled from the posterior probability distribution (BA) were used. For the latter, 180 000 post burn-in trees were thinned to obtain 1000 trees in Burntrees v. 0.1.9 (Nylander 2011). Polytomies were randomly resolved in the APE module (Paradis et al. 2004) in R (R Development Core Team 2010). Because the variable describing parasitism in our system has 3 levels (free-living, parasitic as deutonymph, permanently parasitic, see below), reconstructions were performed under a 6-rate model (transition rates between all character states are different). The single-rate model (all transition rates are equal) was also evaluated to see if it better fits the data. BA were run under reversible jump MCMC, where models are visited in proportion to their posterior probabilities. An exponential prior was used seeded from a uniform on the interval 0–30.

We tested the hypothesis of irreversibility of parasitism (ecological interpretation of Dollo's law—no morphological characters subject to Dollo's law could be found) using a recently developed framework (Goldberg and Igic 2008) in DiversiTree containing implementations of binary state speciation and extinction (BiSSE) (Maddison et al. 2007), multiple state speciation and extinction (MuSSE) (FitzJohn 2010), and Markov models of character state evolution (mk2 and mk3) (Pagel 1994). We used 3 possible schemes of character state coding: (i) 3-state coding (free-living, parasitic as deutonymphs, and parasitic in all stages) and 2 approaches of binary coding (ii–iii) where the “parasitic as deutonymphs” state was assigned to either “free-living” or to “parasitic,” respectively. Models corresponding to the 3 coding schemes were constrained to conform to the assumptions of irreversible evolution (root is fixed to the free-living state; back transition rates are set to zero) and then compared to unconstrained models using the AIC. For the unconstrained models, stationary probabilities based on the assumption of equilibrium in the state frequencies at the root were used. The importance of estimating diversification rates can be demonstrated by the following example. If diversification of free-living mites (state A) is greater than that of parasitic mites (state B), then the B-to-A transition rate would be overestimated by simple mk models leading to an incorrect rejection of irreversible evolution. If, however, diversification is independent from state A, then mk models can be used to assess unidirectional evolution. For these reasons, we give results of different analyses estimating state transitions rates only (mk) or both transition and diversification rates (BiSSE and MuSSE). Irreversible evolution is rejected when all tests prefer unconstrained models over models constrained for irreversible evolution with AIC differences of more than 10 (Burnham and Anderson 2004).

Estimating diversification rates, as in the above tests, may be affected by incomplete or biased taxonomic sampling (Cusimano and Renner 2010). Our data set is incomplete, but the taxon sampling is nearly unbiased (see the section “Taxon Sampling, DNA Isolation, and Sequencing”). Consequently, results obtained by the BiSSE and MuSSE analyses should be treated with caution.

Estimating Host Specificity

Parasites associated with a single host, or closely related hosts, are more likely to be specialists, while those parasitizing multiple, distantly related hosts are expected to be generalists. Generalists would not have adaptations to a particular host taxon but, having a general morphology and means of active dispersal to different hosts, rather employ the strategy of being jacks-of-all-trades (Krasnov et al. 2004). In contrast to highly specialized parasites, they have a greater evolutionary potential and are more likely to be involved in a possible transition from a parasitic to free-living state.

To estimate host specificity in Psoroptidia, we prepared a host–parasite database consolidating data provided by H. Proctor (bird mites) (Proctor 2012) and A. Bochkov (mammal mites, unpublished). All records involving unidentified mites or hosts were excluded. The MS Excel function vlookup was used to verify mite names using the Index to Organism Names compiled from Zoological Record® (Thomson Reuters 2012); host names were similarly verified and standardized using the online databases Mammal Species of the World (Wilson and Reeder 2005) and the Clements Checklist of Birds of the World (Clements et al. 2011). After these procedures, our database contained a total of 9483 unique host records for 3901 named species of psoroptidian mites (excluding Pyroglyphidae). These data were then used to extract host range values (expressed as the number of host taxa per mite species) at 4 taxonomical levels (species-, genus-, family-, and order-specific) using Pivot Table summary functions in Excel. Averaged values were also calculated for each family of mites. Parasites were arbitrarily considered as “generalists” if they were associated with multiple host orders, while species- to family-specific parasites were considered as “specialists.”

Results

Phylogenetic Analyses

Our analyses recovered pyroglyphids deeply nested within a large monophyletic lineage otherwise comprising parasites of mammals and birds—an unranked clade termed Psoroptidia (BS 100, PP 1.00; Fig. 1a: node 3). This group was first proposed by Murray (1877), and named so by Yunker (1955), but its monophyly has been contested multiple times. OConnor (1982) offered convincing morphological evidence for a monophyletic Psoroptidia and also proposed its rank-free nature. Monophyly was confirmed by DNA sequence analyses (Klimov and OConnor 2008). The family Heterocoptidae, which is parasitic on insects, is consistently recovered as the sister group of Psoroptidia (BS 100, PP 1.00; Fig. 1a: node 2). The internal lineage including the dust mites, the EP complex (Klimov and OConnor 2008) (BS 99, PP 1.00; Fig. 1a: node 4) was inferred as an assemblage of bird nasal endoparasites (family Turbinoptidae: Congocoptes, Schoutedenocoptes), bird skin mites (family Epidermoptidae: Microlichus, Myialges, Promyialges, the latter 2 genera are also partially hyperparasitic on insects), feather mites (family Psoroptoididae: Picalgoides, Mesalgoides, Hyomesalges, Temnalges), and mammal skin mites (family Psoroptidae: Otodectes, Psoroptes and family Lobalgidae: Echimytricalges). Three key nodes (Fig. 1a: nodes 2–4) render the dust mites deeply nested within parasitic lineages and are particularly strongly supported by different types of total evidence, as well as independent data partition, and alternative close and distant outgroup analyses (Table 1; Supplementary Figs. S3 and S4). For the total evidence analyses support was high (BS 99–100), decreasing for analyses including only rDNA (BS 96–100) or protein-coding genes (BS 69–89) (Table 1). Single-gene analyses also recovered these clades, but often with low or no support (Table 1; Supplementary Fig. S3).

Table 1.

Support for 4 key nodes leading to the dust mites under different analytical techniques, total evidence, and independent data partition analyses

Partition Node
 
 
All genes (BA) 0.99 1.00 1.00 1.00 
All genes (ML) 46 100 100 99 
rDNA 31a 100 98 96 
Protein-coding genes 44 89 73 69 
18S nr 94 96 44 
28S 1b 99 96 57 
EF1-α 0b 6c 6c 0d 
SRP54 8a 8c 8c 52 
HSP70 7a 38 39 
Partition Node
 
 
All genes (BA) 0.99 1.00 1.00 1.00 
All genes (ML) 46 100 100 99 
rDNA 31a 100 98 96 
Protein-coding genes 44 89 73 69 
18S nr 94 96 44 
28S 1b 99 96 57 
EF1-α 0b 6c 6c 0d 
SRP54 8a 8c 8c 52 
HSP70 7a 38 39 

Notes: Node 1, Hypoderatidae+Heterocoptidae+Psoroptidia; Node 2, Heterocoptidae+Psoroptidia; Node 3, Psoroptidia; Node 4, EP complex (Fig. 1). Support is indicated as Bayesian posterior probabilities (BA) and nonparametric bootstrap support values (ML analysis; for brevity, ML is not indicated in rows 3–7); nr = not recovered.

aAlso includes Suidasiidae.

bAlso includes several families of free-living mites.

cWith Heterocoptidae as ingroup.

dRecovered monophyletic, except for Gymnoglyphus.

In agreement with other studies (Fain 1962; Cui et al. 2009; Arlian and Morgan 2011), we note paraphyly of the large pyroglyphid genus Dermatophagoides. The total evidence tree (Fig. 1a) and individual analyses of 4 genes (Supplementary Fig. S3) show that the free-living pyroglyphids are not monophyletic. However, one gene, HSP70, does render them monophyletic (Fig. 1b), suggesting that reversal to a free-living state might have occurred only once. To further investigate the discordance across gene genealogies, we estimated a species tree focusing on the free-living pyroglyphid clades using the coalescent-based species tree method. Free-living pyroglyphids were monophyletic (PP 0.49) on a species tree inferred from a subset of the 3 protein-coding genes entered into analysis as DNA (Fig. 1c). Two gene trees (EF1-α and HSP70) supported these relationships (PP 0.54–0.67), while another gene (SRP54) also included Onychalges (bird parasites) here as an ingroup (PP 1.00) (Supplementary Fig. S5). Translating these loci into protein, adding rDNA, or using rDNA only resulted in nonmonophyly of the free-living dust mites (Supplementary Fig. S5). Given these results, a single origin of the free-living lifestyle in pyroglyphids cannot be ruled out. However, to accurately estimate the probability of this event, as well as internal relationships of Pyroglyphidae, a larger number of protein-coding loci should be used.

Alternative Relationships of the Dust Mites

We statistically tested 62 existing hypotheses of dust mite phylogenetic affinities based on morphological, molecular, and immunological evidence and rejected 57 of them. Although several hypotheses may be of interest to mite taxonomists only, we note a subset of 16 hypotheses that assume close relationships of pyroglyphids to the storage mites (Supplementary Table S6). These relationships were proposed chiefly based on the large, superficially similar female ovipore and associated structures in the free-living family Chortoglyphidae and the dust mites (Berlese 1897; Cunliffe 1958). With the replacement of chortoglyphids by other nonpsoroptodian Astigmata, this idea re-emerged in several recent molecular studies (Loo et al. 2003; Suarez-Martinez et al. 2005; Dabert et al. 2010; Yang et al. 2010). The controversy also extends to immunology, with different studies demonstrating similar cross-reactivity of the dust mites and either storage mites or parasitic psoroptidian mites (reviewed Sidenius et al. 2001; Ong and Chew 2010). Given that the mites mostly have species-specific allergens (Arlian et al. 2009), one would expect higher positive covariation between sensitization involving phylogenetically related groups. All hypotheses involving close relationships of the storage and dust mites were statistically rejected by our Consel analyses.

The 5 hypotheses that cannot be rejected by our analyses suggest a parasitic sister group of the dust mites. These hypotheses indicate (i) the possibility of close relationships of the free-living pyroglyphids (Dermatophagoidinae+Pyroglyphinae), and, thus, a single origin of the free-living style (Supplementary Table S6, row 22); (ii) mammal parasites (family Psoroptidae) being their sister group (Supplementary Table S6, row 6); (iii) the removal of Ptyssalgidae (hummingbird quill inhabitants) from the house dust mite complex with placement within the pteronyssid feather mites (Supplementary Table S6, row 39); and (iv–v) the higher-level relationships of Psoroptidia with the 4 critical nodes as shown on Figure 1a (Supplementary Table S6, rows 1, 52). Hypotheses i and ii (with the inclusion of Turbinoptidae) were only recovered in the analysis utilizing species tree estimation under the coalescent model (Fig. 1c), whereas other supported hypotheses were consistently recovered on our concatenated phylogeny (Fig. 1a) and all other types of analyses.

Irreversible Evolution

Our analyses indicate that unconstrained models perform substantially better than models constrained for unidirectional evolution (the root is fixed to the free-living state, the transition rate of parasitism-to-free-living state is set to zero) in all pairwise comparisons representing different coding strategies for the variable “parasitism.” The full-rate models (Table 2, rows 3, 4, 9, 10, 15, 16) were preferred over the single-rate models (rows 1, 2, 7, 8, 13, 14), justifying the use of the former over the transition-only models (mk). In these best, unconstrained, full-rate models, speciation rates were inferred to be much higher, and the corresponding extinction rates were estimated to be much lower for the parasitic state than the free-living state (Table 2, rows 3, 9, 15). In contrast, under the assumptions of irreversible evolution, the increase in diversification rates should be associated with the free-living state. No differences in results were observed in 3 separate analyses utilizing different coding strategies for mites with parasitic deutonymphs but otherwise free living [i.e., assigned to a separate category (Table 2, rows 1–6), or classified as either free living (rows 7–12) or parasitic (rows 13–18)].

Table 2.

Comparison of models constrained for irreversible evolution (Dollo's law) of parasitism vs. unconstrained models

States Model Root λλλμμμq12 q13 q21 q23 q31 q32 Ln L df AIC ΔAIC 
MuSSE Stationary  7.060  1.322   0.000 0.014 3.312 0.536 0.122 0.000 218.095 −420.190 40.789 
  Fixed  7.060  1.322   0.170 0.513 — 0.000 — — 202.889 −395.777 65.202 
  Stationary 4.878 7.749 8.629 3.365 0.000 0.000 0.000 0.000 6.107 0.516 0.152 0.000 242.490 12 −460.979 
  Fixed 4.832 7.393 8.808 1.041 7.208 0.000 0.340 0.433 — 0.000 — — 222.772 −427.545 33.434 
 mk3 Stationary — — — — — — 0.143 0.151 0.119 0.096 0.101 0.000 −56.006 124.011 
  Fixed — — — — — — 0.171 0.512 — 0.000 — — −67.568 141.136 17.125 
2a BiSSE Stationary 7.060  — 1.322  — 0.404 — 0.078 — — — 220.632 −433.264 34.342 
  Fixed 7.060  — 1.322  — 0.688 — — — — — 212.007 −418.014 49.591 
  Stationary 4.770 8.546 — 1.812 0.000 — 0.185 — 0.239 — — — 239.803 −467.606 
10   Fixed 4.730 8.635 — 1.067 0.000 — 0.590 — — — — — 229.900 −449.800 17.806 
11  mk2 Stationary — — — — — — 0.404 — 0.078 — — — −49.825 103.649 
12   Fixed — — — — — — 0.688 — — — — — −58.449 118.899 15.250 
13 2b BiSSE Stationary 7.060  — 1.322  — 0.176 — 0.094 — — — 236.047 −464.093 36.190 
14   Fixed 7.060  — 1.321  — 0.487 — — — — — 224.707 −443.415 56.868 
15   Stationary 4.971 8.659 — 1.802 0.000 — 0.127 — 0.125 — — — 256.141 −500.282 
16   Fixed 5.002 8.811 — 1.513 0.000 — 0.397 — — — — — 244.089 −478.177 22.105 
17  mk2 Stationary — — — — — — 0.176 — 0.094 — — — −34.410 72.820 
18   Fixed — — — — — — 0.487 — — — — — −45.749 93.498 20.678 
States Model Root λλλμμμq12 q13 q21 q23 q31 q32 Ln L df AIC ΔAIC 
MuSSE Stationary  7.060  1.322   0.000 0.014 3.312 0.536 0.122 0.000 218.095 −420.190 40.789 
  Fixed  7.060  1.322   0.170 0.513 — 0.000 — — 202.889 −395.777 65.202 
  Stationary 4.878 7.749 8.629 3.365 0.000 0.000 0.000 0.000 6.107 0.516 0.152 0.000 242.490 12 −460.979 
  Fixed 4.832 7.393 8.808 1.041 7.208 0.000 0.340 0.433 — 0.000 — — 222.772 −427.545 33.434 
 mk3 Stationary — — — — — — 0.143 0.151 0.119 0.096 0.101 0.000 −56.006 124.011 
  Fixed — — — — — — 0.171 0.512 — 0.000 — — −67.568 141.136 17.125 
2a BiSSE Stationary 7.060  — 1.322  — 0.404 — 0.078 — — — 220.632 −433.264 34.342 
  Fixed 7.060  — 1.322  — 0.688 — — — — — 212.007 −418.014 49.591 
  Stationary 4.770 8.546 — 1.812 0.000 — 0.185 — 0.239 — — — 239.803 −467.606 
10   Fixed 4.730 8.635 — 1.067 0.000 — 0.590 — — — — — 229.900 −449.800 17.806 
11  mk2 Stationary — — — — — — 0.404 — 0.078 — — — −49.825 103.649 
12   Fixed — — — — — — 0.688 — — — — — −58.449 118.899 15.250 
13 2b BiSSE Stationary 7.060  — 1.322  — 0.176 — 0.094 — — — 236.047 −464.093 36.190 
14   Fixed 7.060  — 1.321  — 0.487 — — — — — 224.707 −443.415 56.868 
15   Stationary 4.971 8.659 — 1.802 0.000 — 0.127 — 0.125 — — — 256.141 −500.282 
16   Fixed 5.002 8.811 — 1.513 0.000 — 0.397 — — — — — 244.089 −478.177 22.105 
17  mk2 Stationary — — — — — — 0.176 — 0.094 — — — −34.410 72.820 
18   Fixed — — — — — — 0.487 — — — — — −45.749 93.498 20.678 

Notes: Analyses are based on selection between models explicitly accounting for speciation (λ), extinction (μ), character state transition rates (q), and character state at root options (root) versus those favoring irreversible evolution (i.e., back character state transitions are set to zero and the root is fixed to the free-living state, root = “fixed,” q21, q31, q32 = “—”). When MuSSE or BiSSE single-rate models (1, 2, 7, 8, 13, 14) are preferred over full-rate models (3, 4, 9, 10, 15, 16) then mk models (5, 6, 11, 12, 17, 18) should be used instead. Here, however, the full-rate models are preferred. Among them, unconstrained models present a better fit to the data (as evidenced by their lowest AICs), thus, rejecting irreversible evolution in this system. States = 3-state coding: free-living, parasitic as deutonymphs, parasitic at all stages; 2a,b = binary coding 1 (categories 1 and 2 joined) and 2 (categories 2 and 3 joined), respectively; Model = BiSSE, MuSSE, mk2-3 (Markov 2 and 3 state models of character state evolution) as implemented in DiversiTree; root = character states at root are either stationary or fixed to state 1 (complex state); λ = speciation rates; μ = extinction rates; q = character state transition rates; ln L = ML; df = number of free parameters (degree of freedom); ΔAIC = models selected by AIC have ΔAICs of zero.

Because unconstrained, full-rate models accounting for diversification and character state transition were preferred over models constrained for irreversible evolution (ΔAIC 33.4, 17.8, 22.1 for the 3 coding strategies; Table 2), and the increase of diversification rates is associated with the parasitic state rather than the free-living state, irreversible evolution is rejected in this system.

Ancestral Ecology

Ancestral character state reconstruction inferred that the respective common ancestors of 2 major lineages leading to the dust mites, Psoroptidia+Heterocoptidae and Psoroptidia (Fig. 1a: nodes 2, 3) were parasitic in all stages of their life cycle (probabilities 0.983–1.000; Supplementary Table S7). Probabilities for ancestral parasitism in the EP complex (Fig. 1a: node 4) were also high (0.684–0.992) and statistically significant (Supplementary Table S7). We note high congruence between 2 independent tests relying on trees inferred by ML and BA. Absolute differences in probabilities yielded by these tests were 0.000–0.154, 0.018±0.033 (range, average ± SD). ML trait models tend to be slightly more conservative than Bayesian models in estimating probabilities for ancestral parasitism (Supplementary Table S7). Constraining the 6-parameter ML trait models to 1 parameter made very little difference to the likelihood (ΔAIC 0.1–0.2, favoring the 6-rate models over 1-rate models). In contrast, Bayesian trait framework using Bayesian trees shows clear superiority of the 1-rate model (ΔAIC 10.7) (Supplementary Table S7).

Specialization, Host Range, and Evolutionary Plasticity

Eighty percent of nonpyroglyphid prosroptidian mites (3109 of 3901 species) are specialists, associated with a single host or closely related hosts belonging to the same genus (Supplementary Fig S8a,b). The large percentage of species with limited host ranges is not surprising because most psoroptidians do not have a specialized dispersal stage, so their main method of infecting new hosts is vertical transfer (from parent to offspring), or, more rarely, horizontal transfer during mating or occasional close contact between potential hosts.

Only 3% (126) of nonpyroglyphid proroptidians are associated with hosts belonging to multiple orders (Supplementary Fig S8a,b). Their ability to live on a range of phylogenetically and morphologically different hosts and maintain gene flow between these seemingly isolated populations suggests that these mites are generalists with potentially higher evolutionary plasticity than single-host mites. On a macroevolutionary scale, such mites are more likely to avoid “dead ends” eminent for single-host parasites with the extinction of their hosts. To successfully colonize multiple hosts, mites should develop (i) “general” morphologies suitable for parasitizing disparate hosts and (ii) means for active dispersal across hosts, including morphological and behavioral adaptations for a short-term “free-living” (or hyperparasitic) period, during which the mites are seeking a new host. Indeed, if psoroptidian families are ordered by their average host range (Supplementary Table S9), the taxa with the broadest host ranges (Psoroptidae, Sarcoptidae, Psoroptoididae, Dermoglyphidae, Knemidokoptidae, Epidermoptidae; host order count per mite species is from 1.122 to 1.767) tend to be skin parasites. This can be explained by the fact that skin, in contrast to feathers or fur, can be similar across many unrelated host lineages, so there is no need for host-specific morphological specializations (such as various elaborate modifications of legs and body to attach to a specific place on the host, as in fur mites, Fig. 2d–f). This is especially true for certain psoroptids parasitizing the mammalian ear—this habitat is relatively well protected from the outside and does not require any specialized structures for fastening to the host. Fur mites and plumaceous feather mites live essentially in a three-dimensional environment, and have very sophisticated morphological adaptations for this. Some feather mites even develop asymmetrical body shapes to match the asymmetry of individual feathers (Gaud and Atyeo 1996). Mammal mites living in upper, dead layers of the epidermis (without a primary contact with the host immune system) are more likely to have broader host range as compared with mites living on the hairs themselves [P = 0.0273–0.0002 for comparisons of species- and genus-specific against order(s)-specific, respectively, our multinomial logistic regression analysis, unpublished].

Several lines of evidence indicate that the EP complex, a group that gave rise to the dust mites (Fig. 1a: node 4), is characterized by a broader host range than most other psoroptidian mites. First, the proportion of mites associated with multiple host orders is relatively higher in the EP complex than in other mites (23.0% vs. 2.3–5.0%) (Supplementary Fig. S8). Second, of the 6 families with the broadest average host range (see above, Supplementary Table S9), 3 (Psoroptidae, Psoroptoididae, and Epidermoptidae) are part of the EP complex. Third, in the EP complex, mammal and bird-associated lineages (Psoroptidae, Lobalgidae vs. Psoroptoididae, Epidermoptidae, Turbinoptidae) appear to be closely related and intermixed (Fig. 1a: node 4). This indicates relatively frequent host shifts across avian and mammalian hosts—something that is not characteristic of other Psoroptidia. Finally, the only family of nonpyroglyphid psoroptidian mites adapted to horizontal transfer belongs to the EP complex (some epidermoptid females are hyperparasitic on hippoboscid flies). These lines of evidence suggest that ancestors of the pyroglyphids were probably unspecialized, multihost mites, parasitizing skin or related environments.

Discussion

Results from our phylogenetic analyses, topology-based tests for alternative placement of house dust mites, ancestral character state reconstructions, and a test for irreversibility of parasitism all suggest that the common ancestor of pyroglyphid house dust mites underwent reversal from a permanently parasitic lifestyle to become secondarily free living. By inference, this violates the ecological interpretation of Dollo's law defining the parasitic state as irreversible because many of the adaptations necessary for life away from the host (“complex state”) are assumed to be lost in parasites (Cruickshank and Paterson 2006). Thus, as exemplified by house dust mites, specialized organisms can adapt to new ecological niches via de-specialization, escaping evolutionary “dead ends.”

Recognizing the possibility of irreversible evolution in this host–parasite system is very important. A pattern of free-living clades deeply nested within parasitic lineages may be generated if free-living lineages speciate faster and go extinct more slowly than parasitic lineages where parasitism is irreversible (Goldberg and Igic 2008). Interestingly, a recently proposed historical ecological scenario for pyroglyphids (Bochkov and Mironov 2011) strikingly resembles this pattern. Under this hypothesis, parasitism has evolved as many as 8 times in Psoroptidia. Pyroglyphids were treated as “living fossils”—the only surviving remnants of the ancestral stem group that gave rise to all of the spectacular diversity of parasitic psoroptidian mites. The authors gave 2 arguments against reversible parasitism: (i) there is no co-divergence of pyroglyphids with their avian hosts and (ii) living on a host versus its nest provides a strong selective advantage. However, strict co-divergence is not expected in this system or elsewhere in Psoroptidia (e.g., their and our cladograms show some mammal-associated mites as the sister group to pyroglyphids). Similarly, the selective advantage argument is not applicable here because mites living on a host versus its nest occupy different ecological niches, so they cannot compete with each other. In addition, because these authors did not use the family Heterocoptidae, a critical parasitic sister group of Psoroptidia, their assumption about the free-living ancestor of psoroptidian mites also should be re-evaluated. Our analyses, accounting for both diversification and character state transition rates, strongly reject the irreversible parasitism scenario in this system. Potentially, these estimates can be confounded by incomplete taxon sampling (Cusimano and Renner 2010), but our tests that do not rely on evaluations of diversification rates arrive at the same conclusion.

How might the ecological shift associated with the reversal to the free-living state have occurred? Close relatives of pyroglyphids were inferred to be mammal skin mites and avian endoparasites [Psoroptidae and Turbinoptidae (Congocoptes)], but more distant relatives might include feather-inhabiting mites (e.g., Analgidae) (Fig. 1a). Although ancestral hosts cannot be determined with certainty due to frequent host shifts, there is little doubt that early free-living dust mites were nest inhabitants. Indeed, nests of birds, mammals, or both are the principal habitat of all modern free-living pyroglyphid species (a few exceptions probably involve accidental records). We propose that a combination of several characteristics of their parasitic ancestors played an important role in abandoning permanent parasitism: tolerance of low humidity, development of powerful digestive enzymes allowing feeding on skin and keratinous materials, and low host specificity with frequent shifts to unrelated hosts. As compared to nonpyroglyphid free-living mites, efficient water balance mechanisms have appeared within parasitic lineages of Psoroptidia (Gaede and Knülle 1987) in response to living in specific regions of the host body (e.g., skin surface, flight feathers) and feeding on a dry, fat-rich diet (keratinaceous materials and lipids of the skin, sebaceous or uropygial gland secretions). The specific diet of the mites and the need to evade the host immune response probably promoted development of powerful, specialized enzymes, such as cysteine protease (group 1 allergen) (Kato et al. 2005). These features, occurring in almost all parasitic mites, were potentially important precursors enabling mite populations to thrive in host nests despite low humidity and scarce, low quality food resources, largely consisting of shed skin and feathers. Furthermore, as evidenced by frequent ancestral shifts to unrelated host lineages, selection in early pyroglyphids appears to have favored an unspecialized morphology (such as the absence of elaborate leg and gnathosomal modifications specialized for attachment to specific places on the hosts, Fig. 2), which probably enabled the switch from permanent parasitism to a completely or partially free-living condition. This transition might have occurred through an intermediate state, with mites living both in the nest and on the body of the host, utilizing the host chiefly for dispersal. This intermediate state can still be seen in several Dermatophagoides species, such as Dermatophagoides evansi (personal observation) and Dermatophagoides pteronyssinus (Fain et al. 1990). With the advent of human civilization, nest-inhabiting pyroglyphids could have shifted to human dwellings from the nests of synanthropic birds or rodents, where, thanks to potent digestive enzymes and other immunogenic molecules, they became a major source of allergies.

From a broader perspective, understanding phylogenetic relationships of house dust mites and other astigmatan mites may provide insights into allergenic properties of their immunogenic proteins and the evolution of paralogs and orthologs of genes encoding allergens. Our phylogenic and historical ecological inferences provide a highly predictive framework for research in immunological aspects of pyroglyphids. Given that the mites mostly have species-specific allergens (Arlian et al. 2009), the great phylogenetic distance between the dust and storage mites might explain the limited cross-reactivity between allergens from these mites demonstrated by some, but not all studies (reviewed Sidenius et al. 2001; Arlian et al. 2009). Similarly, cross-antigenicity between allergens of house dust mites and other parasitic psoroptidians (Arlian et al. 1991) is not surprising given their close phylogenetic affinities. Finally, the ancestrally parasitic lifestyle of the dust mites is a very plausible explanation for the ability of present day dust mites to modulate host immunity and downgrade host immunological response in laboratory settings (Arlian and Morgan 2011). One would expect this feature to be present only in true parasites, having a direct contact with the host immune system.

Supplementary Material

Data files and/or other supplementary information related to this paper have been deposited at Dryad (http://www.datadryad.org/) under doi: 10.5061/dryad.rd1bc.

Funding

This work was supported by National Science Foundation (NSF) [DEB-0613769, 9521744, 0118766 to B.M.OC.], and also benefitted in part from specimens collected by the Field Museum's Emerging Pathogens Project, funded by the Davee Foundation and the Dr. Ralph and Marian Falk Medical Research Trust. The molecular work of this study was conducted in the Genomic Diversity Laboratory of the University of Michigan Museum of Zoology. We greatly appreciate comments and careful proofreading of a previous version of the manuscript by E. Jockusch, J. Wiens, and R. Cruickshank.

Acknowledgments

Specimens for this study were collected by a network of 64 biologists in 19 countries. We thank T. M. Pérez Ortíz, G. Montiel Parra, J. B. Morales-Malacara, L. Luna Wong, W. Stanley, and Z. Sayakova for organizing our fieldtrips and obtaining collecting permits in Mexico, Peru, Tanzania, and Kazakhstan; S. V. Mironov for sharing his unpublished technique of sampling from live birds without euthanizing them; L. Arlian for discussion on allergy issues; J. Hubert for discussion on house dust mite enzymes; J. Brown and H. Lanier for discussion on various issues of molecular phylogenetics; M. Colloff, S. V. Mironov, and A. V. Bochkov for discussion; G. Bauchan and R. Ochoa for taking low temperature scanning electron microscope photographs (LT-SEM) of the American house dust mite Dermatophagoides farinae; H. Abraham, J. Dikema, E. Foot Perkowski, and K. Mar for assistance with molecular laboratory work.

References

Agnarsson
I.
Aviles
L.
Coddington
J.A.
Maddison
W.P.
Sociality in theridiid spiders: repeated origins of an evolutionary dead end
Evolution
 , 
2006
, vol. 
60
 (pg. 
2342
-
2351
)
Altekar
G.
Dwarkadas
S.
Huelsenbeck
J.P.
Ronquist
F.
Parallel Metropolis coupled Markov chain Monte Carlo for Bayesian phylogenetic inference
Bioinformatics
 , 
2004
, vol. 
20
 (pg. 
407
-
415
)
Altschul
S.F.
Madden
T.L.
Schaffer
A.A.
Zhang
J.
Zhang
Z.
Miller
W.
Lipman
D.J.
Gapped BLAST and PSI-BLAST: a new generation of protein database search programs
Nucleic Acids Res.
 , 
1997
, vol. 
25
 (pg. 
3389
-
3402
)
Apakupakul
K.
Siddall
M.E.
Burreson
E.M.
Higher level relationships of leeches (Annelida: Clitellata: Euhirudinea) based on morphology and gene sequences
Mol. Phylogenet. Evol.
 , 
1999
, vol. 
12
 (pg. 
350
-
359
)
Arlian
L.
Morgan
M.
Vyszenski-Moher
D.
Sharra
D.
Cross-reactivity between storage and dust mites and between mites and shrimp
Exp. Appl. Acarol.
 , 
2009
, vol. 
47
 (pg. 
159
-
172
)
Arlian
L.G.
House-dust-mite allergens: a review
Exp. Appl. Acarol.
 , 
1991
, vol. 
10
 (pg. 
167
-
186
)
Arlian
L.G.
Morgan
M.S.
Immunomodulation of skin cytokine secretion by house dust mite extracts
Int. Arch. Allergy Immunol.
 , 
2011
, vol. 
156
 (pg. 
171
-
178
)
Arlian
L.G.
Vyszenski-Moher
D.L.
Ahmed
S.G.
Estes
S.A.
Cross-antigenicity between the scabies mite, Sarcoptes scabiei, and the house dust mite, Dermatophagoides pteronyssinus
J. Invest. Dermatol.
 , 
1991
, vol. 
96
 (pg. 
349
-
354
)
Armbruster
W.S.
Baldwin
B.G.
Switch from specialized to generalized pollination
Nature
 , 
1998
, vol. 
394
 pg. 
632
 
Basagana
X.
Sunyer
J.
Kogevinas
M.
Zock
J.P.
Duran-Tauleria
E.
Jarvis
D.
Burney
P.
Anto
J.M.
Socioeconomic status and asthma prevalence in young adults: the European Community Respiratory Health Survey
Am. J. Epidemiol.
 , 
2004
, vol. 
160
 (pg. 
178
-
188
)
Berlese
A.
Acari Myriopoda et Scorpiones hucusque in Italia reperta. Ordo Cryptostigmata (Sarcoptidae)
 , 
1882–1897
Portici (Italy)
Typis Vesuvianae
Bochkov
A.V.
Mironov
S.V.
Phylogeny and systematics of mammal-associated psoroptidian mites (Acariformes : Astigmata : Psoroptidia) derived from external morphology
Invertebr. Syst.
 , 
2011
, vol. 
25
 (pg. 
22
-
59
)
Borda
E.
Siddall
M.E.
Arhynchobdellida (Annelida: Oligochaeta: Hirudinida): phylogenetic relationships and evolution
Mol. Phylogenet. Evol.
 , 
2004
, vol. 
30
 (pg. 
213
-
225
)
Brown
J.M.
Hedtke
S.M.
Lemmon
A.R.
Lemmon
E.M.
When trees grow too long: investigating the causes of highly inaccurate Bayesian branch-length estimates
Syst. Biol.
 , 
2010
, vol. 
59
 (pg. 
145
-
161
)
Burnham
K.P.
Anderson
D.R.
Multimodel inference—understanding AIC and BIC in model selection
Sociol. Method Res.
 , 
2004
, vol. 
33
 (pg. 
261
-
304
)
Cannone
J.
Subramanian
S.
Schnare
M.
Collett
J.
D'Souza
L.
Du
Y.
Feng
B.
Lin
N.
Madabusi
L.
Muller
K.
Pande
N.
Shang
Z.
Yu
N.
Gutell
R.
The Comparative RNA Web (CRW) Site: an online database of comparative sequence and structure information for ribosomal, intron, and other RNAs
BMC Bioinformatics
 , 
2002
, vol. 
3
 pg. 
2
 
Clements
J.F.
Schulenberg
T.S.
Iliff
M.J.
Sullivan
B.L.
Wood
C.L.
Roberson
D.
The Clements checklist of birds of the world: Version 6.6 [Internet]
2011
 
Colles
A.
Liow
L.H.
Prinzing
A.
Are specialists at risk under environmental change? Neoecological, paleoecological and phylogenetic approaches
Ecol. Lett.
 , 
2009
, vol. 
12
 (pg. 
849
-
863
)
Colloff
M.J.
Dust mites
 , 
2009
Dordrecht (The Netherlands)
Springer
Combes
C.
Simberloff
D.
The art of being a parasite
 , 
2005
Chicago (IL)
University of Chicago Press
Cruickshank
R.H.
Paterson
A.M.
The great escape: do parasites break Dollo's law?
Trends Parasitol.
 , 
2006
, vol. 
22
 (pg. 
509
-
515
)
Cui
Y.
Zhou
P.
Peng
J.
Peng
M.
Zhou
Y.
Lin
Y.
Liu
L.
Cloning, sequence analysis, and expression of cDNA coding for the major house dust mite allergen, Der f 1, in Escherichia coli
Braz. J. Med. Biol. Res.
 , 
2008
, vol. 
41
 (pg. 
380
-
388
)
Cui
Y.B.
Gao
C.X.
Zhou
Y.
Peng
Z.
Peng
M.
Lin
Y.
Peng
J.
Phylogenetic analysis of house dust mites
Cent. Eur. J. Med.
 , 
2009
, vol. 
5
 (pg. 
69
-
74
)
Cunliffe
F.
Pyroglyphus morlani, a new genus and species of mite forming a new family, Pyroglyphidae, in the Acaridiae
Proc. Entomol. Soc. Wash.
 , 
1958
, vol. 
60
 (pg. 
85
-
86
)
Cunnington
A.M.
Gregory
P.H.
Mites in bedroom air
Nature
 , 
1968
, vol. 
217
 (pg. 
1271
-
1272
)
Cusimano
N.
Renner
S.S.
Slowdowns in diversification rates from real phylogenies may not be real
Syst. Biol.
 , 
2010
, vol. 
59
 (pg. 
458
-
464
)
Dabert
M.
Witaliñski
W.
Kazmierski
A.
Olszanowski
Z.
Dabert
J.
Molecular phylogeny of acariform mites (Acari, Arachnida): strong conflict between phylogenetic signal and long-branch attraction artifacts
Mol. Phylogenet. Evol.
 , 
2010
, vol. 
56
 (pg. 
222
-
241
)
Darriba
D.
Taboada
G.L.
Doallo
R.
Posada
D.
HPC selection of models of DNA substitution
2011a
Paris
(pg. 
65
-
72
Proceedings of the 9th International Conference on Computational Methods in Systems Biology (CMSB ‘11), September 2011
Darriba
D.
Taboada
G.L.
Doallo
R.
Posada
D.
ProtTest 3: fast selection of best-fit models of protein evolution
Bioinformatics
 , 
2011b
, vol. 
27
 (pg. 
1164
-
1165
)
Degnan
J.H.
Rosenberg
N.A.
Gene tree discordance, phylogenetic inference and the multispecies coalescent
Trends Ecol. Evol.
 , 
2009
, vol. 
24
 (pg. 
332
-
340
)
D'Haese
C.
Is psammophily an evolutionary dead end? A phylogenetic test in the genus Willemia (Collembola : Hypogastruridae)
Cladistics
 , 
2000
, vol. 
16
 (pg. 
255
-
273
)
Dieterich
C.
Sommer
R.J.
How to become a parasite—lessons from the genomes of nematodes
Trends Genet.
 , 
2009
, vol. 
25
 (pg. 
203
-
209
)
Domes
K.
Althammer
M.
Norton
R.A.
Scheu
S.
Maraun
M.
The phylogenetic relationship between Astigmata and Oribatida (Acari) as indicated by molecular markers
Exp. Appl. Acarol.
 , 
2007
, vol. 
42
 (pg. 
159
-
171
)
Dorris
M.
Viney
M.E.
Blaxter
M.L.
Molecular phylogenetic analysis of the genus Strongyloides and related nematodes
Int. J. Parasitol.
 , 
2002
, vol. 
32
 (pg. 
1507
-
1517
)
Drummond
A.J.
Rambaut
A.
BEAST: Bayesian evolutionary analysis by sampling trees
BMC Evol. Biol.
 , 
2007
, vol. 
7
 pg. 
214
 
Dunn
C.W.
Hejnol
A.
Matus
D.Q.
Pang
K.
Browne
W.E.
Smith
S.A.
Seaver
E.
Rouse
G.W.
Obst
M.
Edgecombe
G.D.
Sorensen
M.V.
Haddock
S.H.D.
Schmidt-Rhaesa
A.
Okusu
A.
Kristensen
R.M.
Wheeler
W.C.
Martindale
M.Q.
Giribet
G.
Broad phylogenomic sampling improves resolution of the animal tree of life
Nature
 , 
2008
, vol. 
452
 (pg. 
745
-
U745
)
Fain
A.
Un acarien remarquable combinant les caracteres de plusieurs families: Evansacarus lari n.g., n
sp. Bull. Ann. Soc. R. Belge Entomol.
 , 
1962
, vol. 
98
 (pg. 
125
-
145
)
Fain
A.
Guerin
B.
Hart
B.J.
Mites and allergic disease
 , 
1990
Varennes an Argonne, France
Allerbio
FitzJohn
R.G.
Quantitative traits and diversification
Syst. Biol.
 , 
2010
, vol. 
59
 (pg. 
619
-
633
)
Futuyma
D.J.
Moreno
G.
The evolution of ecological specialization
Annu. Rev. Ecol. Syst.
 , 
1988
, vol. 
19
 (pg. 
207
-
233
)
Gaede
K.
Knülle
W.
Water vapour uptake from the atmosphere and critical equilibrium humidity of a feather mite
Exp. Appl. Acarol.
 , 
1987
, vol. 
3
 (pg. 
45
-
52
)
Gaud
J.
Atyeo
W.T.
Feather mites of the World (Acarina, Astigmata): the supraspecific taxa. Part 1. Text
Koninklijk Museum voor Midden Afrika Tervuren Belgie Annalen Zoologische Wetenschappen
 , 
1996
, vol. 
277
 (pg. 
1
-
193
)
Gillespie
J.J.
Johnston
J.S.
Cannone
J.J.
Gutell
R.R.
Characteristics of the nuclear (18S, 5.8S, 28S and 5S) and mitochondrial (12S and 16S) rRNA genes of Apis mellifera (Insecta: Hymenoptera): structure, organization, and retrotransposable elements
Insect Mol. Biol.
 , 
2006
, vol. 
15
 (pg. 
657
-
686
)
Goldberg
E.E.
Igic
B.
On phylogenetic tests of irreversible evolution
Evolution
 , 
2008
, vol. 
62
 (pg. 
2727
-
2741
)
Hammad
H.
Chieppa
M.
Perros
F.
Willart
M.A.
Germain
R.N.
Lambrecht
B.N.
House dust mite allergen induces asthma via Toll-like receptor 4 triggering of airway structural cells
Nat. Med.
 , 
2009
, vol. 
15
 (pg. 
410
-
416
)
Heled
J.
Drummond
A.J.
Bayesian inference of species trees from multilocus data
Mol. Biol. Evol.
 , 
2010
, vol. 
27
 (pg. 
570
-
580
)
Holmes
E.B.
Is specialization a dead end
Am. Nat.
 , 
1977
, vol. 
111
 (pg. 
1021
-
1026
)
Holt
P.G.
Thomas
W.R.
Sensitization to airborne environmental allergens: unresolved issues
Nat. Immunol.
 , 
2005
, vol. 
6
 (pg. 
957
-
960
)
Janz
N.
Nyblom
K.
Nylin
S.
Evolutionary dynamics of host-plant specialization: a case study of the tribe Nymphalini
Evolution
 , 
2001
, vol. 
55
 (pg. 
783
-
796
)
Kato
T.
Takai
T.
Mitsuishi
K.
Okumura
K.
Ogawa
H.
Cystatin A inhibits IL-8 production by keratinocytes stimulated with Der p 1 and Der f 1: biochemical skin barrier against mite cysteine proteases
J. Allergy Clin. Immunol.
 , 
2005
, vol. 
116
 (pg. 
169
-
176
)
Klimov
P.B.
OConnor
B.M.
Origin and higher-level relationships of psoroptidian mites (Acari: Astigmata: Psoroptidia): evidence from three nuclear genes
Mol. Phylogenet. Evol.
 , 
2008
, vol. 
47
 (pg. 
1135
-
1156
)
Knowles
L.
Klimov
P.B.
Estimating phylogenetic relationships despite discordant gene trees across loci: the species tree of a diverse species group of feather mites (Acari: Proctophyllodidae)
Parasitology
 , 
2011
, vol. 
138
 (pg. 
1750
-
1759
)
Kocot
K.M.
Cannon
J.T.
Todt
C.
Citarella
M.R.
Kohn
A.B.
Meyer
A.
Santos
S.R.
Schander
C.
Moroz
L.L.
Lieb
B.
Halanych
K.M.
Phylogenomics reveals deep molluscan relationships
Nature
 , 
2011
, vol. 
477
 (pg. 
452
-
456
)
Kohlsdorf
T.
Lynch
V.J.
Rodrigues
M.T.
Brandley
M.C.
Wagner
G.P.
Data and data Interpretation in the study of limb evolution: a reply to Galis et al. On the reevolution of digits in the lizard genus Bachia
Evolution
 , 
2010
, vol. 
64
 (pg. 
2477
-
2485
)
Krasnov
B.R.
Poulin
R.
Shenbrot
G.I.
Mouillot
D.
Khokhlova
I.S.
Ectoparasitic “jacks-of-all-trades”: relationship between abundance and host specificity in fleas (Siphonaptera) parasitic on small mammals
Am. Nat.
 , 
2004
, vol. 
164
 (pg. 
506
-
516
)
Lanyon
S.M.
Interspecific brood parasitism in blackbirds (Icterinae)—a phylogenetic perspective
Science
 , 
1992
, vol. 
255
 (pg. 
77
-
79
)
Lee
K.H.
Yu
H.S.
Park
S.K.
Lee
S.J.
Lee
K.A.
Kim
S.M.
Ock
M.S.
Jeong
H.J.
Phylogeny of mite taxa (Acari:Sarcoptiformes) based on small subunit ribosomal RNA sequences
J. Life Sci.
 , 
2006
, vol. 
16
 (pg. 
71
-
75
)
Light
J.E.
Siddall
M.E.
Phylogeny of the leech family Glossiphoniidae based on mitochondrial gene sequences and morphological data
J. Parasitol.
 , 
1999
, vol. 
85
 (pg. 
815
-
823
)
Lloyd
C.M.
Dust mites' dirty dealings in the lung
Nat. Med.
 , 
2009
, vol. 
15
 (pg. 
366
-
367
)
Loo
A.H.B.
Tan
S.P.L.
Angus
A.C.
Kuay
K.T.
Reginald
K.
Gao
Y.F.
Chew
F.T.
Genetic relationship between allergy-causing dust mites: phylogenetic inference from random amplified polymorphic DNA (RAPD) markers, housekeeping gene (18S rDNA), and group 2 allergens
J. Allergy Clin. Immunol.
 , 
2003
, vol. 
111
 pg. 
S162
 
Lynch
V.J.
Wagner
G.P.
Did egg-laying boas break Dollo's law? Phylogenetic evidence for reversal to oviparity in sand boas (Eryx: Boidae)
Evolution
 , 
2010
, vol. 
64
 (pg. 
207
-
216
)
Maddison
W.P.
Midford
P.E.
Otto
S.P.
Estimating a binary character's effect on speciation and extinction
Syst. Biol.
 , 
2007
, vol. 
56
 (pg. 
701
-
710
)
Marshall
D.C.
Cryptic failure of partitioned Bayesian phylogenetic analyses: lost in the land of long trees
Syst. Biol.
 , 
2010
, vol. 
59
 (pg. 
108
-
117
)
Mathews
D.H.
Sabina
J.
Zuker
M.
Turner
D.H.
Expanded sequence dependence of thermodynamic parameters improves prediction of RNA secondary structure
J. Mol. Biol.
 , 
1999
, vol. 
288
 (pg. 
911
-
940
)
Mendonca
A.G.
Alves
R.J.
Pereira-Leal
J.B.
Loss of genetic redundancy in reductive genome evolution
Plos Comput. Biol.
 , 
2011
, vol. 
7
 (pg. 
1
-
10
)
Morse
G.E.
Farrell
B.D.
Ecological and evolutionary diversification of the seed beetle genus Stator (Coleoptera : Chrysomelidae : Bruchinae)
Evolution
 , 
2005
, vol. 
59
 (pg. 
1315
-
1333
)
Murray
A.
Economic Entomology, Aptera: prepared at the request of the Lords of the Committee of Council on education, and published for them
 , 
1877
London
Chapman and Hall
Norton
R.A.
Morphological evidence for the evolutionary origin of Astigmata (Acari: Acariformes)
Exp. Appl. Acarol.
 , 
1998
, vol. 
22
 (pg. 
559
-
594
)
Nosil
P.
Mooers
A.O.
Testing hypotheses about ecological specialization using phylogenetic trees
Evolution
 , 
2005
, vol. 
59
 (pg. 
2256
-
2263
)
Nylander
J.A.
Wilgenbusch
J.C.
Warren
D.L.
Swofford
D.L.
AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics
Bioinformatics
 , 
2008
, vol. 
24
 (pg. 
581
-
583
)
Nylander
J.A.A.
Burntrees v. 0.1.9. Perl script for manipulating MrBayes tree and parameter files. Program distributed by the author
2011
Uppsala (Sweden)
Evolutionary Biology Centre, Uppsala University
OConnor
B.M.
Parker
S.P.
Astigmata
Synopsis and classification of living organisms
 , 
1982
New York
McGraw-Hill Book Company
(pg. 
146
-
169
)
Ong
S.T.
Chew
F.T.
Pawankar
R.
Holgate
S.T.
Rosenwasser
L.J.
Reconstructing the repertoire of mite allergens by recombinant DNA technology
Allergy frontiers: future perspectives
 , 
2010
Japan
Springer
(pg. 
49
-
64
)
Pagel
M.
Detecting correlated evolution on phylogenies: a general method for the comparative analysis of discrete characters
Proc. R. Soc. Lond. B Biol. Sci.
 , 
1994
, vol. 
255
 (pg. 
37
-
45
)
Pagel
M.
Meade
A.
Barker
D.
Bayesian estimation of ancestral character states on phylogenies
Syst. Biol.
 , 
2004
, vol. 
53
 (pg. 
673
-
684
)
Paradis
E.
Claude
J.
Strimmer
K.
APE: analyses of phylogenetics and evolution in R language
Bioinformatics
 , 
2004
, vol. 
20
 (pg. 
289
-
290
)
Parfrey
L.W.
Grant
J.
Tekle
Y.I.
Lasek-Nesselquist
E.
Morrison
H.G.
Sogin
M.L.
Patterson
D.J.
Katz
L.A.
Broadly sampled multigene analyses yield a well-resolved eukaryotic tree of life
Syst. Biol.
 , 
2010
, vol. 
59
 (pg. 
518
-
533
)
Pepato
A.R.
da Rocha
C.E.
Dunlop
J.A.
Phylogenetic position of the acariform mites: sensitivity to homology assessment under total evidence
BMC Evol. Biol.
 , 
2010
, vol. 
10
 pg. 
235
 
Prendini
L.
Francke
O.F.
Vignoli
V.
Troglomorphism, trichobothriotaxy and typhlochactid phylogeny (Scorpiones, Chactoidea): more evidence that troglobitism is not an evolutionary dead-end
Cladistics
 , 
2010
, vol. 
26
 (pg. 
117
-
142
)
Price
P.W.
Evolutionary biology of parasites
 , 
1980
Princeton (NJ)
Princeton University Press
Proctor
H.C.
Feather mites (Acari: Astigmata): ecology, behavior and evolution
Annu. Rev. Entomol.
 , 
2003
, vol. 
48
 (pg. 
185
-
209
)
Proctor
H.C.
List of feather mites and their hosts [Internet]
2012
 
R Development Core Team
R: a language and environment for statistical computing, reference index version 2.11.1 [Internet]
 , 
2010
Vienna (Austria)
R Foundation for Statistical Computing
 
ISBN 3-900051-07-0. Available from: URL http://www.R-project.org
Radovsky
F.J.
Krantz
G.W.
Whitaker
J.O.
A remarkable example of predation in the parasitic mite family Macronyssidae
Int. J. Acarol.
 , 
1997
, vol. 
23
 (pg. 
3
-
6
)
Rambaut
A.
Drummond
A.J.
Tracer v1.5 [Internet]
2009
 
Ronquist
F.
Huelsenbeck
J.P.
MRBAYES 3: Bayesian phylogenetic inference under mixed models
Bioinformatics
 , 
2003
, vol. 
19
 (pg. 
1572
-
1574
)
Rothfels
C.J.
Larsson
A.
Kuo
L.Y.
Korall
P.
Chiou
W.L.
Pryer
K.M.
Overcoming deep roots, fast rates, and short internodes to resolve the ancient rapid radiation of eupolypod II ferns
Syst. Biol.
 , 
2012
, vol. 
61
 (pg. 
490
-
509
)
Sakharkar
K.R.
Dhar
P.K.
Chow
V.T.K.
Genome reduction in prokaryotic obligatory intracellular parasites of humans: a comparative analysis
Int. J. Syst. Evol. Microbiol.
 , 
2004
, vol. 
54
 (pg. 
1937
-
1941
)
Salewski
V.
Satellite species in lampreys: a worldwide trend for ecological speciation in sympatry?
J. Fish Biol.
 , 
2003
, vol. 
63
 (pg. 
267
-
279
)
Shimodaira
H.
Multiple comparisons of log-likelihoods and combining nonnested models with applications to phylogenetic tree selection
Commun. Stat. Theory Methods
 , 
2001
, vol. 
30
 (pg. 
1751
-
1772
)
Shimodaira
H.
An approximately unbiased test of phylogenetic tree selection
Syst. Biol.
 , 
2002
, vol. 
51
 (pg. 
492
-
508
)
Siddall
M.E.
Brooks
D.R.
Desser
S.S.
Phylogeny and the reversibility of parasitism
Evolution
 , 
1993
, vol. 
47
 (pg. 
308
-
313
)
Sidenius
K.E.
Hallas
T.E.
Poulsen
L.K.
Mosbech
H.
Allergen cross-reactivity between house-dust mites and other invertebrates
Allergy
 , 
2001
, vol. 
56
 (pg. 
723
-
733
)
Smith
B.P.
Loss of larval parasitism in parasitengonine mites
Exp. Appl. Acarol.
 , 
1998
, vol. 
22
 (pg. 
187
-
199
)
Stireman
J.O.
The evolution of generalization? Parasitoid flies and the perils of inferring host range evolution from phylogenies
J. Evol. Biol.
 , 
2005
, vol. 
18
 (pg. 
325
-
336
)
Struck
T.H.
Paul
C.
Hill
N.
Hartmann
S.
Hosel
C.
Kube
M.
Lieb
B.
Meyer
A.
Tiedemann
R.
Purschke
G.
Bleidorn
C.
Phylogenomic analyses unravel annelid evolution
Nature
 , 
2011
, vol. 
471
 (pg. 
95
-
98
)
Suarez-Martinez
E.B.
Montealegre
F.
Sierra-Montes
J.M.
Herrera
R.J.
Molecular identification of pathogenic house dust mites using 12S rRNA sequences
Electrophoresis
 , 
2005
, vol. 
26
 (pg. 
2927
-
2934
)
Sukumaran
J.
Holder
M.T.
DendroPy: a Python library for phylogenetic computing
Bioinformatics
 , 
2010
, vol. 
26
 (pg. 
1569
-
1571
)
Takebayashi
N.
Morrell
P.L.
Is self-fertilization an evolutionary dead end? Revisiting an old hypothesis with genetic theories and a macroevolutionary approach
Am. J. Bot.
 , 
2001
, vol. 
88
 (pg. 
1143
-
1150
)
Termonia
A.
Hsiao
T.H.
Pasteels
J.M.
Milinkovitch
M.C.
Feeding specialization and host-derived chemical defense in chrysomeline leaf beetles did not lead to an evolutionary dead end
Proc. Natl. Acad. Sci. U. S. A.
 , 
2001
, vol. 
98
 (pg. 
3909
-
3914
)
Thomson Reuters
Index to Organism Names (ION) [Internet]
2012
 
Available from: URL http://www.organismnames.com/
Toft
C.A.
Aeschlimann
A.
Bolis
L.
Parasite-host associations: coexistence or conflict?
 , 
1991
New York
Oxford University Press
Tovey
E.R.
Chapman
M.D.
Platts-Mills
T.A.
Mite faeces are a major source of house dust allergens
Nature
 , 
1981
, vol. 
289
 (pg. 
592
-
593
)
Wiens
J.J.
Re-evolution of lost mandibular teeth in frogs after more than 200 million years, and re-evaluating Dollo's law
Evolution
 , 
2011
, vol. 
65
 (pg. 
1283
-
1296
)
Wilson
D.E.
Reeder
D.M.
Mammal species of the world. A taxonomic and geographic reference
2005
3rd ed
Baltimore (MD)
The Johns Hopkins University Press
Yang
B.
Cai
J.
Cheng
X.
Identification of astigmatid mites using ITS2 and COI regions
Parasitol. Res.
 , 
2010
, vol. 
108
 (pg. 
497
-
503
)
Yunker
C.E.
A proposed classification of the Acaridiae (Acarina: Sarcoptiformes)
Proc. Helminthol. Soc. Wash.
 , 
1955
, vol. 
22
 (pg. 
98
-
105
)
Zuker
M.
Mfold web server for nucleic acid folding and hybridization prediction
Nucleic Acids Res.
 , 
2003
, vol. 
31
 (pg. 
3406
-
3415
)
Zwickl
D.J.
Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion [Ph.D. dissertation]
 , 
2006
[Austin (TX)]
The University of Texas at Austin

Author notes

Associate Editor: Elizabeth Jockusch