Next Article in Journal
Benzoic Acid and Its Hydroxylated Derivatives Suppress Early Blight of Tomato (Alternaria solani) via the Induction of Salicylic Acid Biosynthesis and Enzymatic and Nonenzymatic Antioxidant Defense Machinery
Next Article in Special Issue
Inactivation of Dermatophytes Causing Onychomycosis Using Non-Thermal Plasma as a Prerequisite for Therapy
Previous Article in Journal
Smuts to the Power of Three: Biotechnology, Biotrophy, and Basic Biology
Previous Article in Special Issue
Comparative Efficacy of Antifungal Agents Used in the Treatment of Oropharyngeal Candidiasis among HIV-Infected Adults: A Systematic Review and Network Meta-Analysis
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
JoF
Volume 7
Issue 8
10.3390/jof7080661
jof-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Accounting for the Biological Complexity of Pathogenic Fungi in Phylogenetic Dating

by
Hannah M. Edwards
* and
Johanna Rhodes
MRC Centre for Global Infectious Disease Analysis, Imperial College School of Public Health, Imperial College London, Norfolk Place, London W2 1PG, UK
*
Author to whom correspondence should be addressed.
J. Fungi 2021, 7(8), 661; https://doi.org/10.3390/jof7080661
Submission received: 4 July 2021 / Revised: 11 August 2021 / Accepted: 11 August 2021 / Published: 14 August 2021
(This article belongs to the Special Issue Fungal Pathogens and Human Health)
Browse Figures
Review Reports Versions Notes

Abstract

:
In the study of pathogen evolution, temporal dating of phylogenies provides information on when species and lineages may have diverged in the past. When combined with spatial and epidemiological data in phylodynamic models, these dated phylogenies can also help infer where and when outbreaks occurred, how pathogens may have spread to new geographic locations and/or niches, and how virulence or drug resistance has developed over time. Although widely applied to viruses and, increasingly, to bacterial pathogen outbreaks, phylogenetic dating is yet to be widely used in the study of pathogenic fungi. Fungi are complex organisms with several biological processes that could present issues with appropriate inference of phylogenies, clock rates, and divergence times, including high levels of recombination and slower mutation rates although with potentially high levels of mutation rate variation. Here, we discuss some of the key methodological challenges in accurate phylogeny reconstruction for fungi in the context of the temporal analyses conducted to date and make recommendations for future dating studies to aid development of a best practices roadmap in light of the increasing threat of fungal outbreaks and antifungal drug resistance worldwide.

1. Introduction

Phylogenetic inference can be a powerful tool to understand the evolutionary history of a pathogen, including species divergence, time to most recent common ancestors and, combined with spatial data, when it may have emerged in a new geographic region. When phylodynamic models are also applied, certain processes that have shaped this evolution can be inferred, such as temporal shifts in selection pressure, changes in effective population size, and spatial dynamics of the host population. Insights can therefore be gathered on how the epidemiology and biology of pathogens have been shaped by this ecological and evolutionary history [1,2].
Scaling of divergence times in phylogenetics makes use of estimates of molecular clock rates. Sampling dates are used to fix the ages of the tips in the phylogenetic tree and estimation of molecular clock rates enables dating of internal nodes corresponding to coalescent or birth events where lineages diverged from their common ancestor [3,4]. The ability to do this is dependent on how the rates of evolutionary and ecological changes relate to each other and there is a requirement for a high amount of detectable genetic variation over the time scale from which the genomes being analysed have been isolated [5]. Hence, in the study of microbial pathogens, these analyses have been limited largely to viruses and some bacteria whose rate of evolutionary change can be detectable over the course of an outbreak. For example, estimates of the mutation rate of ebolavirus from the 2014 West African epidemic suggest a mutation rate of ~1.2 × 10−3 [1.13, 1.27 × 10−3] nucleotide substitutions (subs) per site per year [6], while Zika virus has been estimated to range from 0.6 × 10−3 [0.3, 0.8 × 10−3] to 0.8 × 10−3 [0.5, 1.1 × 10−3] subs/site/year [7]. Temporal analyses have been applied less to pathogenic fungi, which typically evolve much slower, although some do evolve within the same range of bacteria to which dating analyses have been applied more frequently. Bacterial mutation rates span several orders of magnitude from ~10−5 to 10−8 subs/site/year [8] including one of the most commonly analysed bacteria, Mycobacterium tuberculosis (MTB), which evolves an estimated 0.5 subs/genome/year, or 1.1 × 10−7 subs/site/year given its 4.4 Mb genome [9]. In comparison, our review of the published literature identified primary dating analysis in just four studies on pathogenic fungi, including Cryptococcus gattii [10], Paracoccidioides spp. [11], and Candida auris [12,13], estimating mutation rates in the order of 10−8, 10−9, and 10−5 subs/site/year for the three pathogens, respectively. It is worth noting, however, that these four species do not represent all fungal phyla with pathogenic members.
Despite limited focus so far, application of temporal analyses such as these is increasingly relevant given fungal outbreaks are increasing in frequency and scope [14,15,16,17,18]. Scaling of divergence times in a phylogeny from genetic distance into absolute units of time can aid in the investigation of pathogen outbreaks—including timing when and from where an outbreak occurred and how it spread—transmission events, and spread of drug resistant clades [2,5,19,20]. Population dynamics can also be timed with historic events, such as human migrations [21], political events [22,23], or public health control measures [24,25]. The yeast C. auris, which causes an invasive blood stream infection, has been the cause of several recent nosocomial outbreaks in contaminated settings with subsequent transmission between patients [26,27,28]. Accurately recreating emergence and transmission events can help identify likely sources of introduction and how such hospital outbreaks have occurred. Several pathogenic fungi have also shown increased resistance to frontline drugs, including Aspergillus fumigatus [29], C. auris [30,31], and Cryptococcus neoformans [32,33], while others have shown recent emergence in global regions previously free of the pathogen and subsequent disease, for example, the emergence of C. gattii in the Pacific Northwest [34,35]. Being able to understand and date these evolutionary histories is key to understanding the processes that shape resistance emergence and spread to new global populations and patient cohorts.
Fungi are complex eukaryotic organisms, with several biological processes that may hinder accurate phylogenetic reconstruction and temporal dating. Fungal mutation rates are slower than those of viruses and (most) bacteria, which may impede detection of a temporal signal among collected isolates, and they have complex reproductive processes leading to potentially high levels of recombination and genetic variation. Fungi also display rate heterogeneity. Multiple species have been shown to display hypermutation, an excess of mutations throughout the genome, which would artificially inflate any clock rate discerned, as well as the same species entering dormant states where mutation rates may be reduced. Here, we discuss the methodological implications of these aspects of fungal biology on temporal analyses, and how they can be addressed in the context of the four primary pathogenic fungal dating studies we have identified to date (herein referred to by their first author: Chow [13], Rhodes [12], Roe [10] and Munoz [11]). These studies utilise differing methodologies with various strengths and deficits to each approach. We end with some recommendations to guide methodological choices in future temporal studies on pathogenic fungi.

2. Methodological Considerations for Temporal Analyses of Pathogenic Fungi

2.1. Detection of a Temporal Signal in Sequence Data

Viruses and bacteria lend themselves well to phylogenetic inference because of their relatively high mutation rates, in the order of 10−8 to 10−6 substitutions per nucleotide site per cell infection (s/n/c) for DNA viruses and 10−6 to 10−4 s/n/c or higher for RNA viruses, and 10−5 to 10−8 s/site/year in bacteria [8,36,37,38]. However, these can often be higher among outbreak lineages (see rate variation discussion below). Therefore, over the course of an outbreak, isolates often show enough genetic variation to enable estimation of the molecular clock. The presence of temporal signal in a set of isolates is an essential pre-requisite for subsequent phylogenetic dating analysis. Sampled isolates have to be spread over an adequate time frame such that the substitution rate can be accurately estimated as there has to be a measurable amount of evolutionary change [39].
Temporal signal can be assessed via a linear regression of sampling time (date of isolation) against root-to-tip distance in the tree (e.g., number of nucleotide substitutions), where the slope corresponds to the substitution rate assuming a strict molecular clock, and where R2 is a measure of how much the evolution shows a temporal signal [40] (Figure 1). This on its own, however, does not assess whether the signal is significantly different to what may be observed by chance. An alternative/additional method to determine whether there is a clock-like signal is to compare models with and without the inclusion of sampling dates. If the addition of sampling dates improves the model fit, then this may be indicative of temporal signal in the data [41,42]. These standard tests for temporal signal can still be insufficient, though. The rates and timescale can be spurious if the number of samples is small, if the sampling period is too short, or if closely related sequences are more likely to have been sampled at similar times [41].
Another method of validating rates and timescales is to use a date-randomisation test using replicate datasets in which sampling times have been randomised [43]. If the estimated mean substitution rate of the data in question lies outside of the 95% credible intervals of the rate estimates calculated from the replicate datasets, then it is indicative of a significant temporal signal. In tests, however, this can fail to detect rate estimates from data with no temporal signal, and it can also perform poorly when sampling times are not uniformly distributed through the tree [44]. As such, a more conservative criterion is recommended, whereby the 95% credible intervals of both the mean estimate from the data and the estimates provided by the randomised replicates do not overlap [44]. Despite this, the randomised permutation approach may still be inappropriate if the genetic and temporal data are confounded such as when closely related sequences are sampled at the same time. In this case, a clustered permutation approach has been proposed [41,44].
Fungi tend to mutate at slower rates than viruses and bacteria, often in the range of 10−8 to 10−10 substitutions per base pair per generation [45,46]. This may affect the ability to detect a temporal signal, particularly if the isolates being analysed have been collected over a short timeframe. Of the four studies on pathogenic fungi reviewed here, three tested their data for evidence of temporal signal (Roe, Rhodes, and Chow), and one (Munoz) did not. C. auris is an intrinsically drug-resistant nosocomial pathogen. Population genomic studies have identified a strong phylogenetic structure containing four clades representing geographical regions: South Asia (Clade I), East Asia (Clade II), Africa (Clade III), and South America (Clade IV). Each Clade is separated by tens of thousands of single-nucleotide polymorphisms (SNPs). The studies investigating C. auris (Chow and Rhodes) each found mutation rates in the order of 10−5 for their C. auris clades, a higher mutation rate than is typical of fungi, and thus perhaps enabling detection of a temporal signal, despite a narrow window of time within which isolates were collected. However, each study only conducted linear regression analyses to determine the temporal signal. Chow identified R2 values of 0.55, 0.56, and 0.21 for C. auris Clades I, III, and IV, respectively, while Rhodes found an R2 of 0.37 for Clade I. In each instance, these were assessed to be indicative of sufficient clock-like signal to justify subsequent dating and did not do further analysis to prove the significance of the temporal signal.
Roe’s study on C. gattii also utilised regression analysis but went further to determine the statistical significance of these measures by conducting 10,000 random date permutations for each of the clades. For the three sub-populations of C. gattii analysed, they identified R2 values of 0.5971, 0.661, and 0.0745 for clades VGIIa, VGIIb, and VGIIc, respectively, indicative of a strong clock-like signal for VGIIa and VGIIb and weak signal for VGIIc. However, the random permutations did not find any of these to be significant from chance at the 0.05 confidence level, with derived p-values of 0.076, 0.268, and 0.294, respectively. Despite the lack of evidence for a statistically significant temporal signal in the data, the authors continued with their dating analysis, and this has been adopted to theorise about how and when C. gattii may have emerged in the Pacific Northwest (PNW) [47].
It should be a priority for future fungal dating studies to rigorously test for a temporal signal in their data and employ statistically significant cut-offs for these signals. Erroneous assumption of a temporal signal would lead to incorrect estimation of the molecular clock rate and incorrect dating of the phylogeny.

2.2. Recombination Needs to Be Considered and Accounted for

Pathogenic fungi, such as Cryptococcus and Candida, propagate predominantly via asexual budding, whereby haploid cells undergo mitosis to produce clonal haploid progeny [48]. However, these fungi also occur as one of two mating types that can reproduce sexually (albeit infrequently), generating novel phenotypes [49,50]. Populations of pathogenic fungi involved in outbreaks generally appear highly clonal; however, these outbreak lineages typically stem from environmental populations with a much greater pool of diversity where evidence of sexual reproduction can be found. This ability to switch between reproductive modes complicates the fungal life cycle and allows for the emergence of new and, occasionally, more virulent genotypes. For example, a clonal population of C. gattii VGIIa was the major lineage involved in the Vancouver and PNW outbreak of cryptococcosis and emerged with a hypervirulent phenotype from less virulent progenitors through a combination of microevolution, recombination, and sexual reproduction [50].
Fungal sexual reproduction can be heterothallic when two opposing mating types come into contact and mate with each other, or homothallic, whereby two cells of a single mating type mate with each other [51]. This ability to undergo unisexual reproduction is important because heterothallic reproduction is likely very rare. In Cryptococcus spp., for example, the MATα mating type is highly prevalent, but the MATa mating type is rarely found in the environment [52,53], while in Candida albicans, heterothallic reproduction is infrequent because it is regulated by the phenotypic switching of MATα or MATa cells into a form that mates more efficiently [54,55]. The same phenotypic switching is not a prerequisite for homothallic mating. Although offspring from unisexual crosses have lower average recombination rates than those derived from bisexual crosses, they do result in divergent and aneuploid progeny able to undergo subsequent meiosis and sporulation [56]. Further complicating the matter, Saccharomyces cerevisiae, for example, can undergo rare matings, which are difficult to distinguish from homothallic mating. These rare matings occur when diploid cells heterozygous for mating type and which thus lack mating ability (these cells are referred to as “non-maters”) convert to a homozygous mating type and cross with maters to produce polyploid progeny [57]. Sexual reproduction of diploid C. albicans cells is also followed by a non-meiotic process of depolyploidization known as concerted chromosome loss (CCL) to convert tetraploid cell products to a diploid state. This CCL is associated with a three-fold greater rate of recombination than normal mitotic growth [58].
These reproductive states present an issue for temporal analysis through the introduction of potentially high amounts of recombination. Recombination has been shown to disrupt phylogenetic inference through loss of any temporal signal and overestimation of substitution rate heterogeneity, leading to loss of the molecular clock, incorrect branch lengths, and erroneous TMRCA estimates (Figure 2) [42,59,60]. The estimated tree topology can be unreliable and lead to a false inference of positive selection [61,62,63]. Ancestral trees become more “star-like” with longer terminal branches, resembling a phylogeny for which a population is undergoing exponential growth (Figure 3). Lengthened branches analogous to an exponentially growing population occur because recombination makes distances between sequences more similar to each other, although this is dependent on the relatedness of the sequences involved in the recombination event and on the time when the event occurred [59,63,64]. Furthermore, ancestral sequence reconstruction is biased by recombination and can be quite distinct from the actual most recent common ancestor (MRCA), instead resembling a concatenate of partial MRCAs at each recombination fragment [65].
Generally, removal of recombinant regions is advised prior to phylogeny estimation. Comparison of Bayesian dating algorithms between trees uncorrected and corrected for recombination has shown that correcting the tree will identify significant temporal signal between isolates otherwise lost in uncorrected trees and will greatly reduce uncertainty in node and root dates [42]. Despite this, the dating papers we identified for pathogenic fungi have failed to adequately account for recombination in their data. In Chow’s study of C. auris, the authors found no evidence of admixture between the four major clades by principal component analysis and genome-wide FST despite each clade being entirely made up of one mating type, potentially allowing for heterothallic reproduction between clades. However, no test for recombination was applied prior to their temporal dating analysis to confirm this, nor to detect recombination sites from reproduction within clades, the latter of which may be more likely since this does not require phenotypic switching. Rhodes identified one isolate that showed less genetic divergence from the root than expected in their root-to-tip divergence analysis, which the authors propose to be possibly due to excessive recombination in this isolate. They did not, however, test isolates for recombination prior to conducting this temporal analysis.
Roe did not find any evidence of recombination between isolates within each of their three sub-population clades of C. gattii (VGIIa, VGIIb, and VGIIc); however, they did not consider potential recombination between these clades. These C. gattii lineages are highly clonal and isolates have been exclusively of one mating type; however, unisexual mating may occur. Finally, Munoz found evidence of recombination between lineages of P. brasiliensis, but they did not account for this in their phylogenetic analysis.
Future studies should seek to adequately consider and identify recombination within their data and remove any identified areas from analysis as appropriate. Tools such as Gubbins [66] and ClonalFrameML [67] offer computationally quick methods with which to identify areas of recombination and accurately reconstruct phylogenies from whole genome sequence data. Gubbins can be applied to a variety of haploid genotypes and large datasets of closely related sequences, although accuracy can reduce when recombination events involve only a small number of bases (the default threshold is set to 3 bp) or when divergence between sequenced isolates increases. Despite these limitations, small recombination events are unlikely to have a large impact on the structure of the inferred genealogy, and the effect of greater sequence divergence can be mitigated, to some extent, by denser sampling or subdivision of the population prior to analysis [66]. These two tools identify recent recombination events between closely related isolates only. A newer tool, fastGEAR, can detect both recent and ancestral recombinations among species-wide gene alignments; however, it does not build a phylogeny [68]. Choice of the correct tool may, therefore, depend on the data and question at hand. Furthermore, these tools have been developed for haploid bacterial genomes and their application to fungi appears largely non-existent in the published literature. A comparison of software tools relevant for dating phylogenies is presented in Table 1. This list is by no means exhaustive but suggests some of the more relevant and popular tools in use. Additional tools relevant to some of the tasks can be found online at: http://methodspopgen.com (accessed on 9 August 2021) [69].

2.3. Quiescence to Hypermutation—Potential Impact of Mutation Rate Variation

Similar to some pathogenic bacteria, fungi can show enormous variation in their metabolic and reproductive rates when subjected to external stressors, resulting in slower or faster rates of mutation over time. Best understood and observed is the emergence of hypermutative states in fungi, such as Cryptococcus and Candida, during infection and exposure to the host environment and anti-fungal drug treatment, leading to emergence of resistant isolates and persistent and/or recurrent infection [45,70,71,72,73,74]. Similar phenomenon is seen in some bacteria, such as Salmonella enterica and Escherichia coli, which have up to a 1000-fold increase in point mutations in clinical isolates [75,76].
Being difficult to observe in the lab, the process whereby fungal cells enter states of latency, dormancy, and/or quiescence is less well understood [77]. A period of latent infection is suspected in many cryptococcal cases, where infection may occur months-to-years before reactivation and manifestation of disease [78,79,80]. The exact length of this period is currently unknown but may be up to 10 years or more in some cases [81]. The dimorphic fungus Talaromyces marneffei (formerly known as Penicillium marneffei) can cause a latent infection spanning several decades, reactivating if the patient later becomes immunosuppressed [82]. A similar process is observed in TB, where MTB cells lie dormant in lung and extrapulmonary tissues [83]. Latent tuberculosis (TB) was traditionally thought to comprise a population of nonreplicating and metabolically inactive cells, but there has since been conflicting evidence about whether this is true or not [83]. Some studies have suggested replication continues as normal in latent TB cell populations [84], while others suggest there is a significant slowing down [85,86,87]. A more recent study investigating index TB cases and paired household contacts that developed active TB up to 5.25 years later has suggested a decline in mutation rates 1–2 years after transmission [88].
As well as these latent states of infection, it has been hypothesised that environmental fungi may enter long periods of dormancy (defined by an absence of metabolic activity such as that observed in fungal spores) or quiescence (where metabolic activity is ongoing and cells are able to return to the “normal” cell cycle) in the environment when resources may be scarce [89,90,91,92]. For example, Saccharomyces cerevisiae has been shown to enter a dormant state when there is lack of sufficient nitrogen or carbon in its substrate [90]. Previous phylogenetic analysis of cryptococcal isolates in Vietnam showed long terminal branches which could indicate a period of cessation of recombination reflective of a possible quiescent state [93]. It is unknown how often these periods of quiescence occur and how long they last, and it may be difficult to determine their impact on dating estimates. Whereas hypermutation is likely to elevate the molecular clock estimate, one could assume that quiescent and dormant states would make clock rates appear slower than the “normal” rate in these isolates. Isolates under study are most often taken from active infection, and thus, if the potentially elevated clock rate is applied across the whole tree, this could estimate incorrect branching times if the internal branches include latent/dormant states where the mutation rate should be much slower. The impact of quiescence could be more complex, given that this state has been associated with its own “replication-independent” mutational profile in response to external stress in laboratory settings, with more deletions relative to insertions and a reduction in A/T composition—features, which are the antithesis to normal cell cycling conditions [94].
If variation in mutation rates is frequent, then it is unlikely a reliable clock rate can be estimated. Temporal heterogeneity has previously been highlighted as a potential issue for both bacteria and viruses since it may not be represented using current clock models and may even obscure the signal of measurable evolution [95]. At the very least, the presence of such states suggests a relaxed clock model to be optimal to take into account at least some variation in the rate of molecular evolution between lineages and allow non-clock-like relationships between sequences within a phylogeny [96,97,98].
Depending on the data, variations of the relaxed clock model, such as the fixed local clock, whereby the underlying tree has a constant rate, but the rate of each clade is able to differ from this global rate [99]; the uncorrelated relaxed clock, whereby each branch of the tree is allowed its own mutation rate independent of the rate on neighbouring branches [98]; or the random local clock, which forms an interim between the strict and relaxed clocks, where each branch can take a different rate or remain the same as one another [100]. More recently, additive clock models have been proposed that satisfy the additive property often unsatisfied by conventional relaxed clock models [101] and may in most cases be more appropriate than the conventional relaxed clock. In the fungal studies identified, both C. auris papers (Chow and Rhodes) utilised strict clock models, seemingly without comparison with alternative options. In their study on Paracoccidioides, Munoz used a relaxed clock/uncorrelated lognormal clock for genome-wide variable sites and strict clock for nonpartitioned variant sites. Finally, Roe tested a combination of relaxed lognormal and strict clock models and chose the best fitting model for each clade, which were strict clock models for VGIIb and VGIIc and a relaxed lognormal model for VGIIa.
More work is needed to understand when and how often fungi may enter these different states and what effect this may have on dating estimates. In the meantime, any analyses should be interpreted with caution and alternative clock models considered.

2.4. Potential Sensitivity to Priors Requires Explicit Sharing of Bayesian Model Settings

A widely recommended tool for phylogenetic and dating analysis is BEAST (Bayesian estimation analysis for sampling trees) which uses Bayesian Markov chain Monte Carlo (MCMC) models to estimate phylogenies and the associated parameters, including effective population size, TMRCAs, and clock rates [102]. Although many studies have suggested Bayesian analysis may be more powerful than comparative maximum likelihood methods, they can be complicated to implement since they require a high amount of computational power as well as knowledge of prior information to inform suitable model parameters, including priors on the tree model, clock rate, population size, substitution model, and partitioning of variant and non-variant sites [103]. Over-parameterisation of the models can result in inference difficulties and/or computational problems, while under-parameterization can lead to incorrect phylogenetic trees and biased estimates of branch lengths and substitution rates with spuriously high posterior probabilities and over-confident uncertainty estimates [104].
Setting correct priors is essential as certain estimates can be sensitive to the priors placed on them [105]. TB, for example, has been shown to be sensitive to both the clock prior and tree prior if the temporal signal is weak in the data set [37]. The approach to partitioning of substitution rates can also affect tree topology, branch lengths and bootstrap support, and divergence time estimation [106,107]. As such, as well as robustly measuring for temporal signal (as described above), it is essential to provide information on the priors set in such analyses, so that any biases can be identified and ensure findings can be replicated and updated as necessary. For this reason, publication of the XML files used for any BEAST analysis is strongly encouraged although appears to be rarely done in practice [102]. In our identified studies, three made use of Bayesian dating analysis using BEAST. None of these studies provides the full model details or XML file used, and the level of detail in the relevant Methods section varies. Furthermore, it is unclear whether two (Munoz and Roe) of the studies removed burn-in from the MCMC algorithm, even though this is a necessary methodological step.

3. Recommendations for the Way Forward

There is no doubt that with the increasing incidence of fungal outbreaks, spread of anti-fungal drug resistance, and appearance of fungal strains in new geographic locations, temporal dating analyses, as applied in phylodynamic models, will be an invaluable tool to understanding these processes and helping to curb the health impact. Although generally slower, fungal mutation rates do overlap with those of bacteria, and outbreak isolates may qualify as “measurably evolving”, and thus applicable to dating methods, particularly as sequencing and analytical tools continue to evolve more power and accuracy. However, the field of phylogenetic dating in fungal pathogens is still in its infancy with differing strengths and deficits in approaches taken to date. For future phylodynamic studies in fungi, some of the lessons described here can be applied to aid development of a “best practice” roadmap, which can be improved upon over time as new tools and techniques become available. A reasonable starting point is to ensure the following steps are addressed (as described in Figure 4):
  • Identify and remove regions of recombination, including recombination both within and between sub-populations/clades. This can be done using freely available tools, such as Gubbins or ClonalFrameML, and should be conducted first as recombination can hinder detection of a temporal signal in the data.
  • Test data for temporal signal through a combination of methods including root-to-tip regression, data randomisation and improved model fit. Importantly, if the data does not have measurable temporal signals, then dating analyses should not be applied.
  • Consider the biology of the organism and potential rate heterogeneity in the selection of appropriate clock models, particularly if datasets are limited to just clinical outbreak isolates or if common ancestors are far back in time. Tests of model fit between different clock (and demographic) models can have poor performance and should not be solely relied upon [108]. Since most fungal infections are saprophytic, stemming from the environment, inclusion of environmental fungal isolates in a phylogeny may provide better indication of heterogeneity in clock rates. Further research into fungal states of hypermutation and quiescence may shed more light on these variations and how frequent or infrequent they may be.
  • If utilising Bayesian analysis, there should be a biological justification for prior settings, and all should be explicitly shared in publications. This can ensure that models are replicable and would allow for comparison between different models. Since final trees can be affected by sensitivity to certain priors, being explicit with model settings would allow for this to be investigated. Studies could examine the sensitivity of the final tree to clock rate and tree priors to justify (or question) the validity of the final tree.

Author Contributions

Conceptualization, H.M.E. and J.R.; writing—original draft preparation, H.M.E.; writing—review and editing, H.M.E. and J.R. All authors have read and agreed to the published version of the manuscript.

Funding

H.M.E. acknowledges funding from Imperial College London and the MRC Centre for Global Infectious Disease Analysis (reference MR/R015600/1), jointly funded by the UK Medical Research Council (MRC) and the UK Foreign, Commonwealth & Development Office (FCDO), under the MRC/FCDO Concordat agreement and is also part of the EDCTP2 programme supported by the European Union. J.R. is supported by a grant from the Wellcome Trust (219551/Z/19/Z).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Grenfell, B.T.; Pybus, O.; Gog, J.; Wood, J.L.N.; Daly, J.; Mumford, J.A.; Holmes, E. Unifying the Epidemiological and Evolutionary Dynamics of Pathogens. Science 2004, 303, 327–332. [Google Scholar] [CrossRef] [Green Version]
  2. Volz, E.M.; Koelle, K.; Bedford, T. Viral Phylodynamics. PLoS Comput. Biol. 2013, 9, e1002947. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  3. Kumar, S.; Hedges, S.B. Advances in Time Estimation Methods for Molecular Data. Mol. Biol. Evol. 2016, 33, 863–869. [Google Scholar] [CrossRef] [Green Version]
  4. Boskova, V.; Bonhoeffer, S.; Stadler, T. Inference of Epidemiological Dynamics Based on Simulated Phylogenies Using Birth-Death and Coalescent Models. PLoS Comput. Biol. 2014, 10, e1003913. [Google Scholar] [CrossRef] [Green Version]
  5. Pybus, O.G.; Rambaut, A. Evolutionary analysis of the dynamics of viral infectious disease. Nat. Rev. Genet. 2009, 10, 540–550. [Google Scholar] [CrossRef] [PubMed]
  6. Holmes, E.C.; Dudas, G.; Rambaut, A.; Andersen, K.G. The evolution of Ebola virus: Insights from the 2013–2016 epidemic. Nature 2016, 538, 193–200. [Google Scholar] [CrossRef] [Green Version]
  7. Barzilai, L.P.; Schrago, C.G. The range of sampling times affects Zika virus evolutionary rates and divergence times. Arch. Virol. 2019, 164, 3027–3034. [Google Scholar] [CrossRef]
  8. Duchene, S.; Holt, K.; Weill, F.-X.; Le Hello, S.; Hawkey, J.; Edwards, D.J.; Fourment, M.; Holmes, E.C. Genome-scale rates of evolutionary change in bacteria. Microb. Genom. 2016, 2, e000094. [Google Scholar] [CrossRef]
  9. Walker, T.M.; Ip, C.L.; Harrell, R.H.; Evans, J.T.; Kapatai, G.; Dedicoat, M.J.; Eyre, D.; Wilson, D.; Hawkey, P.M.; Crook, D.W.; et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: A retrospective observational study. Lancet Infect. Dis. 2013, 13, 137–146. [Google Scholar] [CrossRef] [Green Version]
  10. Roe, C.C.; Bowers, J.; Oltean, H.; DeBess, E.; Dufresne, P.J.; McBurney, S.; Overy, D.P.; Wanke, B.; Lysen, C.; Chiller, T.; et al. Dating the Cryptococcus gattii Dispersal to the North American Pacific Northwest. mSphere 2018, 3, e00499-17. [Google Scholar] [CrossRef] [Green Version]
  11. Muñoz, J.; Farrer, R.A.; Desjardins, C.A.; Gallo, J.E.; Sykes, S.; Sakthikumar, S.; Misas, E.; Whiston, E.A.; Bagagli, E.; Soares, C.M.A.; et al. Genome Diversity, Recombination, and Virulence across the Major Lineages of Paracoccidioides. mSphere 2016, 1, e00213-16. [Google Scholar] [CrossRef] [Green Version]
  12. Rhodes, J.; Abdolrasouli, A.; Farrer, R.A.; Cuomo, C.A.; Aanensen, D.M.; Armstrong-James, D.; Fisher, M.C.; Schelenz, S. Genomic Epidemiology of the UK Outbreak of the Emerging Human Fungal Pathogen Candida auris. Emerg. Microbes Infect. 2018, 7, 43. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  13. Chow, N.A.; Muñoz, J.F.; Gade, L.; Berkow, E.L.; Li, X.; Welsh, R.M.; Forsberg, K.; Lockhart, S.R.; Adam, R.; Alanio, A.; et al. Tracing the Evolutionary History and Global Expansion of Candida auris Using Population Genomic Analyses. mBio 2020, 11, 03364-19. [Google Scholar] [CrossRef] [PubMed]
  14. Fisher, M.C.; Henk, D.A.; Briggs, C.J.; Brownstein, J.S.; Madoff, L.C.; McCraw, S.L.; Gurr, S.J. Emerging fungal threats to animal, plant and ecosystem health. Nature 2012, 484, 186–194. [Google Scholar] [CrossRef]
  15. Garcia-Solache, M.A.; Casadevall, A. Hypothesis: Global Warming Will Bring New Fungal Diseases for Mammals. mBio 2010, 1, 00061-10. [Google Scholar] [CrossRef] [Green Version]
  16. Fisher, M.C.; Gurr, S.J.; Cuomo, C.A.; Blehert, D.S.; Jin, H.; Stukenbrock, E.H.; Stajich, J.E.; Kahmann, R.; Boone, C.; Denning, D.W.; et al. Threats Posed by the Fungal Kingdom to Humans, Wildlife, and Agriculture. mBio 2020, 11, 00449-20. [Google Scholar] [CrossRef]
  17. Benedict, K.; Richardson, M.; Vallabhaneni, S.; Jackson, B.R.; Chiller, T. Emerging issues, challenges, and changing epidemiology of fungal disease outbreaks. Lancet Infect. Dis. 2017, 17, e403–e411. [Google Scholar] [CrossRef]
  18. Bougnoux, M.-E.; Brun, S.; Zahar, J.-R. Healthcare-associated fungal outbreaks: New and uncommon species, New molecular tools for investigation and prevention. Antimicrob. Resist. Infect. Control. 2018, 7, 1–9. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  19. Stadler, T.; Kouyos, R.; von Wyl, V.; Yerly, S.; Böni, J.; Bürgisser, P.; Klimkait, T.; Joos, B.; Rieder, P.; Xie, D.; et al. Estimating the Basic Reproductive Number from Viral Sequence Data. Mol. Biol. Evol. 2012, 29, 347–357. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  20. Lewis, F.; Hughes, G.J.; Rambaut, A.; Pozniak, A.; Brown, A.J.L. Episodic Sexual Transmission of HIV Revealed by Molecular Phylodynamics. PLoS Med. 2008, 5, e50. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  21. Luo, T.; Comas, I.; Luo, D.; Lu, B.; Wu, J.; Wei, L.; Yang, C.; Liu, Q.; Gan, M.; Sun, G.; et al. Southern East Asian origin and coexpansion ofMycobacterium tuberculosisBeijing family with Han Chinese. Proc. Natl. Acad. Sci. USA 2015, 112, 8136–8141. [Google Scholar] [CrossRef] [Green Version]
  22. Eldholm, V.; Pettersson, J.H.-O.; Brynildsrud, O.; Kitchen, A.; Rasmussen, E.M.; Lillebaek, T.; Rønning, J.O.; Crudu, V.; Mengshoel, A.T.; Debech, N.; et al. Armed conflict and population displacement as drivers of the evolution and dispersal of Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 2016, 113, 13881–13886. [Google Scholar] [CrossRef] [Green Version]
  23. Merker, M.; Blin, C.; Mona, S.; Duforet-Frebourg, N.; Lecher, S.; Willery, E.; Blum, M.; Rüsch-Gerdes, S.; Mokrousov, I.; Aleksic, E.; et al. Evolutionary history and global spread of the Mycobacterium tuberculosis Beijing lineage. Nat. Genet. 2015, 47, 242–249. [Google Scholar] [CrossRef]
  24. Dudas, G.; Carvalho, L.M.; Bedford, T.; Tatem, A.J.; Baele, G.; Faria, N.R.; Park, D.J.; Ladner, J.T.; Arias, A.; Asogun, D.; et al. Virus genomes reveal factors that spread and sustained the Ebola epidemic. Nature 2017, 544, 309–315. [Google Scholar] [CrossRef] [Green Version]
  25. Dellicour, S.; Baele, G.; Dudas, G.; Faria, N.R.; Pybus, O.; Suchard, M.A.; Rambaut, A.; Lemey, P. Phylodynamic assessment of intervention strategies for the West African Ebola virus outbreak. Nat. Commun. 2018, 9, 1–9. [Google Scholar] [CrossRef] [Green Version]
  26. Borman, A.M.; Johnson, E.M. Candida auris in the UK: Introduction, dissemination, and control. PLOS Pathog. 2020, 16, e1008563. [Google Scholar] [CrossRef]
  27. Schelenz, S.; Hagen, F.; Rhodes, J.L.; Abdolrasouli, A.; Chowdhary, A.; Hall, A.; Ryan, L.; Shackleton, J.; Trimlett, R.; Meis, J.F.; et al. First hospital outbreak of the globally emerging Candida auris in a European hospital. Antimicrob. Resist. Infect. Control. 2016, 5, 1–7. [Google Scholar] [CrossRef] [Green Version]
  28. Lee, W.G.; Shin, J.H.; Uh, Y.; Kang, M.G.; Kim, S.H.; Park, K.H.; Jang, H.-C. First Three Reported Cases of Nosocomial Fungemia Caused by Candida auris. J. Clin. Microbiol. 2011, 49, 3139–3142. [Google Scholar] [CrossRef] [Green Version]
  29. Toyotome, T.; Hagiwara, D.; Takahashi, H.; Watanabe, A.; Kamei, K. Emerging Antifungal Drug Resistance in Aspergillus fumigatus and Among Other Species of Aspergillus. Curr. Fungal Infect. Rep. 2018, 12, 105–111. [Google Scholar] [CrossRef]
  30. Worku, M.; Girma, F. Candida auris: From Multidrug Resistance to Pan-Resistant Strains. Infect. Drug Resist. 2020, 13, 1287–1294. [Google Scholar] [CrossRef]
  31. Du, H.; Bing, J.; Hu, T.; Ennis, C.L.; Nobile, C.J.; Huang, G. Candida auris: Epidemiology, biology, antifungal resistance, and virulence. PLOS Pathog. 2020, 16, e1008921. [Google Scholar] [CrossRef]
  32. Smith, K.D.; Achan, B.; Hullsiek, K.H.; McDonald, T.R.; Okagaki, L.H.; Alhadab, A.A.; Akampurira, A.; Rhein, J.; Meya, D.B.; Boulware, D.; et al. Increased Antifungal Drug Resistance in Clinical Isolates of Cryptococcus neoformans in Uganda. Antimicrob. Agents Chemother. 2015, 59, 7197–7204. [Google Scholar] [CrossRef] [Green Version]
  33. Zafar, H.; Altamirano, S.; Ballou, E.R.; Nielsen, K. A titanic drug resistance threat in Cryptococcus neoformans. Curr. Opin. Microbiol. 2019, 52, 158–164. [Google Scholar] [CrossRef] [PubMed]
  34. Centers for Disease Control and Prevention (CDC). Emergence of Cryptococcus gattii—Pacific Northwest, 2004–2010. MMWR 2010, 59, 865–868. [Google Scholar]
  35. Bartlett, K.H.; Kidd, S.; Kronstad, J.W. The emergence of Cryptococcus gattii in British Columbia and the Pacific Northwest. Curr. Infect. Dis. Rep. 2008, 10, 58–65. [Google Scholar] [CrossRef] [PubMed]
  36. Peck, K.M.; Lauring, A.S. Complexities of Viral Mutation Rates. J. Virol. 2018, 92, 01031-17. [Google Scholar] [CrossRef] [Green Version]
  37. Menardo, F.; Duchêne, S.; Brites, D.; Gagneux, S. The molecular clock of Mycobacterium tuberculosis. PLOS Pathog. 2019, 15, e1008067. [Google Scholar] [CrossRef] [Green Version]
  38. Duffy, S.; Shackelton, L.A.; Holmes, E. Rates of evolutionary change in viruses: Patterns and determinants. Nat. Rev. Genet. 2008, 9, 267–276. [Google Scholar] [CrossRef] [PubMed]
  39. Drummond, A.; Pybus, O.; Rambaut, A. Inference of Viral Evolutionary Rates from Molecular Sequences. Adv. Parasitol. 2003, 54, 331–358. [Google Scholar] [CrossRef]
  40. Rambaut, A.; Lam, T.T.; Carvalho, L.M.; Pybus, O. Exploring the temporal structure of heterochronous sequences using TempEst (formerly Path-O-Gen). Virus Evol. 2016, 2, vew007. [Google Scholar] [CrossRef] [Green Version]
  41. Murray, G.; Wang, F.; Harrison, E.M.; Paterson, G.; Mather, A.E.; Harris, S.R.; Holmes, M.A.; Rambaut, A.; Welch, J.J. The effect of genetic structure on molecular dating and tests for temporal signal. Methods Ecol. Evol. 2015, 7, 80–89. [Google Scholar] [CrossRef]
  42. Didelot, X.; Croucher, N.; Bentley, S.D.; Harris, S.R.; Wilson, D. Bayesian inference of ancestral dates on bacterial phylogenetic trees. Nucleic Acids Res. 2018, 46, e134. [Google Scholar] [CrossRef] [Green Version]
  43. Ramsden, C.; Melo, F.L.; Figueiredo, L.M.; Holmes, E.C.; Zanotto, P.M.A. High rates of molecular evolution in hantaviruses. Mol. Biol. Evol. 2008, 25, 1488–1492. [Google Scholar] [CrossRef] [Green Version]
  44. Duchene, S.; Duchêne, D.A.; Holmes, E.; Ho, S.Y. The Performance of the Date-Randomization Test in Phylogenetic Analyses of Time-Structured Virus Data. Mol. Biol. Evol. 2015, 32, 1895–1906. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  45. Ene, I.V.; Farrer, R.; Hirakawa, M.; Agwamba, K.; Cuomo, C.A.; Bennett, R.J. Global analysis of mutations driving microevolution of a heterozygous diploid fungal pathogen. Proc. Natl. Acad. Sci. USA 2018, 115, E8688–E8697. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  46. Zhu, Y.O.; Siegal, M.; Hall, D.; Petrov, D.A. Precise estimates of mutation rate and spectrum in yeast. Proc. Natl. Acad. Sci. USA 2014, 111, E2310–E2318. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  47. Engelthaler, D.M.; Casadevall, A. On the Emergence of Cryptococcus gattii in the Pacific Northwest: Ballast Tanks, Tsunamis, and Black Swans. mBio 2019, 10, e02193-19. [Google Scholar] [CrossRef] [Green Version]
  48. Duina, A.A.; Miller, M.E.; Keeney, J.B. Budding Yeast for Budding Geneticists: A Primer on the Saccharomyces cerevisiae Model System. Genetics 2014, 197, 33–48. [Google Scholar] [CrossRef] [Green Version]
  49. Tavanti, A.; Gow, N.A.; Maiden, M.C.; Odds, F.C.; Shaw, D.J. Genetic evidence for recombination in Candida albicans based on haplotype analysis. Fungal Genet. Biol. 2004, 41, 553–562. [Google Scholar] [CrossRef]
  50. Billmyre, R.B.; Croll, D.; Li, W.; Mieczkowski, P.; Carter, D.A.; Cuomo, C.; Kronstad, J.W.; Heitman, J. Highly Recombinant VGII Cryptococcus gattii Population Develops Clonal Outbreak Clusters through both Sexual Macroevolution and Asexual Microevolution. mBio 2014, 5, e01494-14. [Google Scholar] [CrossRef] [Green Version]
  51. Lin, X.; Patel, S.M.; Litvintseva, A.P.; Floyd, A.; Mitchell, T.G.; Heitman, J. Diploids in the Cryptococcus neoformans Serotype A Population Homozygous for the α Mating Type Originate via Unisexual Mating. PLOS Pathog. 2009, 5, e1000283. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  52. Kwon-Chung, K.J.; Bennett, J.E. Distribution of α and α mating types of Cryptococcus neoformans among natural and clinical isolates. Am. J. Epidemiol. 1978, 108, 337–340. [Google Scholar] [CrossRef]
  53. Litvintseva, A.P.; Marra, R.E.; Nielsen, K.; Heitman, J.; Vilgalys, R.; Mitchell, T.G. Evidence of Sexual Recombination among Cryptococcus neoformans Serotype A Isolates in Sub-Saharan Africa. Eukaryot. Cell 2003, 2, 1162–1168. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  54. Bennett, R.J. Coming of Age—Sexual Reproduction in Candida Species. PLOS Pathog. 2010, 6, e1001155. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  55. Miller, M.G.; Johnson, A.D. White-Opaque Switching in Candida albicans Is Controlled by Mating-Type Locus Homeodomain Proteins and Allows Efficient Mating. Cell 2002, 110, 293–302. [Google Scholar] [CrossRef] [Green Version]
  56. Roth, C.; Sun, S.; Billmyre, R.B.; Heitman, J.; Magwene, P.M. A High-Resolution Map of Meiotic Recombination in Cryptococcus deneoformans Demonstrates Decreased Recombination in Unisexual Reproduction. Genetics 2018, 209, 567–578. [Google Scholar] [CrossRef] [Green Version]
  57. Gunge, N.; Nakatomi, Y. Genetic Mechanisms of Rare Matings of the Yeast Saccharomyces cerevisiae Heterozygous for Mating Type. Genetics 1972, 70, 41–58. [Google Scholar] [CrossRef]
  58. Anderson, M.Z.; Thomson, G.J.; Hirakawa, M.P.; Bennett, R.J. A ‘parameiosis’ drives depolyploidization and homologous recombination in Candida albicans. Nat. Commun. 2019, 10, 1–10. [Google Scholar] [CrossRef] [Green Version]
  59. Schierup, M.H.; Hein, J. Consequences of Recombination on Traditional Phylogenetic Analysis. Genetics 2000, 156, 879–891. [Google Scholar] [CrossRef] [PubMed]
  60. Hein, J.; Schierup, M.; Wiuf, C. Gene Genealogies, Variation and Evolution: A Primer in Coalescent Theory; Oxford University Press: Oxford, UK; New York, NY, USA, 2005. [Google Scholar]
  61. 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. [Google Scholar] [CrossRef]
  62. Shriner, D.; Nickle, D.C.; Jensen, M.; Mullins, J.I. Potential impact of recombination on sitewise approaches for detecting positive natural selection. Genet. Res. 2003, 81, 115–121. [Google Scholar] [CrossRef] [Green Version]
  63. Posada, D.; Crandall, K.A. The Effect of Recombination on the Accuracy of Phylogeny Estimation. J. Mol. Evol. 2002, 54, 396–402. [Google Scholar] [CrossRef] [PubMed]
  64. Hedge, J.; Wilson, D.J. Bacterial Phylogenetic Reconstruction from Whole Genomes Is Robust to Recombination but Demographic Inference Is Not. mBio 2014, 5, e02158-14. [Google Scholar] [CrossRef] [Green Version]
  65. Arenas, M.; Posada, D. The Effect of Recombination on the Reconstruction of Ancestral Sequences. Genetics 2010, 184, 1133–1139. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  66. Croucher, N.J.; Page, A.J.; Connor, T.R.; Delaney, A.J.; Keane, J.A.; Bentley, S.D.; Parkhill, J.; Harris, S.R. Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins. Nucleic Acids Res. 2015, 43, e15. [Google Scholar] [CrossRef] [Green Version]
  67. Didelot, X.; Wilson, D.J. ClonalFrameML: Efficient Inference of Recombination in Whole Bacterial Genomes. PLoS Comput. Biol. 2015, 11, e1004041. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  68. Mostowy, R.; Croucher, N.J.; Andam, C.P.; Corander, J.; Hanage, W.P.; Marttinen, P. Efficient Inference of Recent and Ancestral Recombination within Bacterial Populations. Mol. Biol. Evol. 2017, 34, 1167–1182. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  69. Bourgeois, Y.X.C. Going down the rabbit hole: A review on how to link genome-wide data with ecology and evolution in natural populations. bioRxiv 2016, 052761. [Google Scholar] [CrossRef]
  70. Billmyre, R.B.; Clancey, S.A.; Li, L.X.; Doering, T.L.; Heitman, J. 5-fluorocytosine resistance is associated with hypermutation and alterations in capsule biosynthesis in Cryptococcus. Nat. Commun. 2020, 11, 1–9. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  71. Billmyre, R.B.; Clancey, S.A.; Li, L.X.; Doering, T.L.; Heitman, J. Hypermutation in Cryptococcus Reveals a Novel Pathway to 5-Fluorocytosine (5FC) Resistance. bioRxiv 2019, 636928. [Google Scholar] [CrossRef] [Green Version]
  72. Rhodes, J.; Beale, M.A.; Vanhove, M.; Jarvis, J.N.; Kannambath, S.; Simpson, J.A.; Ryan, A.; Meintjes, G.; Harrison, T.; Fisher, M.C.; et al. A Population Genomics Approach to Assessing the Genetic Basis of Within-Host Microevolution Underlying Recurrent Cryptococcal Meningitis Infection. G3 Genes Genomes Genet. 2017, 7, 1165–1176. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  73. Weil, T.; Santamaría, R.; Lee, W.; Rung, J.; Tocci, N.; Abbey, D.; Bezerra, A.; Carreto, L.; Moura, G.R.; Bayés, M.; et al. Adaptive Mistranslation Accelerates the Evolution of Fluconazole Resistance and Induces Major Genomic and Gene Expression Alterations in Candida albicans. mSphere 2017, 2, 00167-17. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  74. Avramovska, O.; Hickman, M.A. The Magnitude of Candida albicans Stress-Induced Genome Instability Results from an Interaction Between Ploidy and Antifungal Drugs. G3 Genes Genomes Genet. 2019, 9, 4019–4027. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  75. LeClerc, J.E.; Li, B.; Payne, W.L.; Cebula, T.A. High Mutation Frequencies among Escherichia coli and Salmonella Pathogens. Sci. Science 1996, 274, 1208–1211. [Google Scholar] [CrossRef] [PubMed]
  76. Le Gall, S.; Desbordes, L.; Gracieux, P.; Saffroy, S.; Bousarghin, L.; Bonnaure-Mallet, M.; Jolivet-Gougeon, A. Distribution of mutation frequencies among Salmonella enterica isolates from animal and human sources and genetic characterization of a Salmonella Heidelberg hypermutator. Vet. Microbiol. 2009, 137, 306–312. [Google Scholar] [CrossRef]
  77. Sagot, I.; Laporte, D. The cell biology of quiescent yeast—A diversity of individual scenarios. J. Cell Sci. 2019, 132, jcs213025. [Google Scholar] [CrossRef] [Green Version]
  78. Salyer, W.R.; Salyer, D.C.; Baker, R.D. Primary Complex of Cryptococcus and Pulmonary Lymph Nodes. J. Infect. Dis. 1974, 130, 74–77. [Google Scholar] [CrossRef]
  79. Goldman, D.L.; Lee, S.C.; Mednick, A.J.; Montella, L.; Casadevall, A. Persistent Cryptococcus neoformans Pulmonary Infection in the Rat Is Associated with Intracellular Parasitism, Decreased Inducible Nitric Oxide Synthase Expression, and Altered Antibody Responsiveness to Cryptococcal Polysaccharide. Infect. Immun. 2000, 68, 832–838. [Google Scholar] [CrossRef] [Green Version]
  80. MacDougall, L.; Fyfe, M. Emergence of Cryptococcus gattii in a Novel Environment Provides Clues to Its Incubation Period. J. Clin. Microbiol. 2006, 44, 1851–1852. [Google Scholar] [CrossRef] [Green Version]
  81. Garcia-Hermoso, D.; Janbon, G.; Dromer, F. Epidemiological Evidence for Dormant Cryptococcus neoformans Infection. J. Clin. Microbiol. 1999, 37, 3204–3209. [Google Scholar] [CrossRef] [Green Version]
  82. Castro-Lainez, M.T.; Sierra-Hoffman, M.; Llompart-Zeno, J.; Adams, R.; Howell, A.; Hoffman-Roberts, H.; Fader, R.; Arroliga, A.C.; Jinadatha, C. Talaromyces marneffei infection in a non-HIV non-endemic population. IDCases 2018, 12, 21–24. [Google Scholar] [CrossRef]
  83. Dutta, N.K.; Karakousis, P.C. Latent Tuberculosis Infection: Myths, Models, and Molecular Mechanisms. Microbiol. Mol. Biol. Rev. 2014, 78, 343–371. [Google Scholar] [CrossRef] [Green Version]
  84. Ford, C.B.; Lin, P.L.; Chase, M.R.; Shah, R.R.; Iartchouk, O.; Galagan, J.; Mohaideen, N.; Ioerger, T.R.; Sacchettini, J.C.; Lipsitch, M.; et al. Use of whole genome sequencing to estimate the mutation rate of Mycobacterium tuberculosis during latent infection. Nat. Genet. 2011, 43, 482–486. [Google Scholar] [CrossRef] [Green Version]
  85. Lillebaek, T.; Dirksen, A.; Baess, I.; Strunge, B.; Thomsen, V.Ø.; Andersen, Å.B. Molecular Evidence of Endogenous Reactivation ofMycobacterium tuberculosisafter 33 Years of Latent Infection. J. Infect. Dis. 2002, 185, 401–404. [Google Scholar] [CrossRef] [Green Version]
  86. Lillebaek, T.; Dirksen, A.; Vynnycky, E.; Baess, I.; Thomsen, V.Ø.; Andersen, Å.B. Stability of DNA Patterns and Evidence ofMycobacterium tuberculosisReactivation Occurring Decades after the Initial Infection. J. Infect. Dis. 2003, 188, 1032–1039. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  87. Colangeli, R.; Arcus, V.L.; Cursons, R.T.; Ruthe, A.; Karalus, N.; Coley, K.; Manning, S.; Kim, S.; Marchiano, E.; Alland, D. Whole Genome Sequencing of Mycobacterium tuberculosis Reveals Slow Growth and Low Mutation Rates during Latent Infections in Humans. PLoS ONE 2014, 9, e91024. [Google Scholar] [CrossRef]
  88. Colangeli, R.; Gupta, A.; Vinhas, S.A.; Venkata, U.D.C.; Kim, S.; Grady, C.; Jones-López, E.C.; Soteropoulos, P.; Palaci, M.; Marques-Rodrigues, P.; et al. Mycobacterium tuberculosis progresses through two phases of latent infection in humans. Nat. Commun. 2020, 11, 1–10. [Google Scholar] [CrossRef]
  89. Roche, B.; Arcangioli, B.; Martienssen, R. Transcriptional reprogramming in cellular quiescence. RNA Biol. 2017, 14, 843–853. [Google Scholar] [CrossRef] [Green Version]
  90. Gray, J.V.; Petsko, G.A.; Johnston, G.C.; Ringe, D.; Singer, R.A.; Werner-Washburne, M. “Sleeping Beauty”: Quiescence in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 2004, 68, 187–206. [Google Scholar] [CrossRef] [Green Version]
  91. De Virgilio, C. The essence of yeast quiescence. FEMS Microbiol. Rev. 2012, 36, 306–339. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  92. Rittershaus, E.S.; Baek, S.-H.; Sassetti, C.M. The Normalcy of Dormancy: Common Themes in Microbial Quiescence. Cell Host Microbe 2013, 13, 643–651. [Google Scholar] [CrossRef] [Green Version]
  93. Ashton, P.M.; Thanh, L.T.; Trieu, P.H.; Van Anh, D.; Trinh, N.M.; Beardsley, J.; Kibengo, F.; Chierakul, W.; Dance, D.A.B.; Rattanavong, S.; et al. Three phylogenetic groups have driven the recent population expansion of Cryptococcus neoformans. Nat. Commun. 2019, 10, 1–10. [Google Scholar] [CrossRef] [Green Version]
  94. Gangloff, S.; Achaz, G.; Francesconi, S.; Villain, A.; Miled, S.; Denis, C.; Arcangioli, B. Quiescence unveils a novel mutational force in fission yeast. eLife 2017, 6, e27469. [Google Scholar] [CrossRef]
  95. Biek, R.; Pybus, O.; Lloyd-Smith, J.O.; Didelot, X. Measurably evolving pathogens in the genomic era. Trends Ecol. Evol. 2015, 30, 306–313. [Google Scholar] [CrossRef] [Green Version]
  96. Thorne, J.L.; Kishino, H.; Painter, I. Estimating the rate of evolution of the rate of molecular evolution. Mol. Biol. Evol. 1998, 15, 1647–1657. [Google Scholar] [CrossRef] [PubMed]
  97. Huelsenbeck, J.P.; Larget, B.; Swofford, D. A Compound Poisson Process for Relaxing the Molecular Clock. Genetics 2000, 154, 1879–1892. [Google Scholar] [CrossRef] [PubMed]
  98. Drummond, A.; Ho, S.Y.W.; Phillips, M.J.; Rambaut, A. Relaxed Phylogenetics and Dating with Confidence. PLoS Biol. 2006, 4, e88. [Google Scholar] [CrossRef] [PubMed]
  99. Yoder, A.D.; Yang, Z. Estimation of Primate Speciation Dates Using Local Molecular Clocks. Mol. Biol. Evol. 2000, 17, 1081–1090. [Google Scholar] [CrossRef] [Green Version]
  100. Drummond, A.J.; Suchard, M.A. Bayesian random local clocks, or one rate to rule them all. BMC Biol. 2010, 8, 114. [Google Scholar] [CrossRef] [Green Version]
  101. Didelot, X.; Siveroni, I.; Volz, E.M. Additive Uncorrelated Relaxed Clock Models for the Dating of Genomic Epidemiology Phylogenies. Mol. Biol. Evol. 2021, 38, 307–317. [Google Scholar] [CrossRef]
  102. Drummond, A.J.; Bouckaert, R.R. BEAST: Bayesian Evolutionary Analysis by Sampling Trees. BMC Evol. Biol. 2007, 7, 214. [Google Scholar] [CrossRef] [Green Version]
  103. Dos Reis, M.; Donoghue, P.; Yang, Z. Bayesian molecular clock dating of species divergences in the genomics era. Nat. Rev. Genet. 2016, 17, 71–80. [Google Scholar] [CrossRef] [Green Version]
  104. Nascimento, F.F.; Dos Reis, M.; Yang, Z. A biologist’s guide to Bayesian phylogenetic analysis. Nat. Ecol. Evol. 2017, 1, 1446–1454. [Google Scholar] [CrossRef] [Green Version]
  105. Möller, S.; du Plessis, L.; Stadler, T. Impact of the tree prior on estimating clock rates during epidemic outbreaks. Proc. Natl. Acad. Sci. USA 2018, 115, 4200–4205. [Google Scholar] [CrossRef] [Green Version]
  106. Angelis, K.; Álvarez-Carretero, S.; Dos Reis, M.; Yang, Z. An Evaluation of Different Partitioning Strategies for Bayesian Estimation of Species Divergence Times. Syst. Biol. 2018, 67, 61–77. [Google Scholar] [CrossRef] [Green Version]
  107. Kainer, D.; Lanfear, R. The Effects of Partitioning on Phylogenetic Inference. Mol. Biol. Evol. 2015, 32, 1611–1627. [Google Scholar] [CrossRef] [Green Version]
  108. Baele, G.; Lemey, P.; Bedford, T.; Rambaut, A.; Suchard, M.A.; Alekseyenko, A. Improving the Accuracy of Demographic and Molecular Clock Model Comparison While Accommodating Phylogenetic Uncertainty. Mol. Biol. Evol. 2012, 29, 2157–2167. [Google Scholar] [CrossRef] [Green Version]
Jof 07 00661 g001 550
Figure 1. Root-to-tip regression analysis to detect evidence of a temporal signal among 20 hypothetical isolates. (A) A phylogenetic tree from which the root-to-tip distance for each isolate is calculated, i.e., the evolutionary change between the time to most recent common ancestor of the sample (TMRCA) and the sampling time of each respective isolate (Tisolate-1, Tisolate-2, Tisolate-n), usually measured as the number of nucleotide substitutions per site over the respective time period. (B) The isolates are plotted based on their date of isolation (x-axis) and root-to-tip divergence, i.e., the number of nucleotide substitutions per site (y-axis). The linear regression provides a measure of temporal signal (R2) as well as molecular clock rate assuming a strict clock is to be applied (slope). In this example, an R2 of 0.432 suggests evidence of a temporal signal in the data, while the (strict) molecular clock rate is estimated at 3.37 × 10−6 subs/site/year.
Figure 1. Root-to-tip regression analysis to detect evidence of a temporal signal among 20 hypothetical isolates. (A) A phylogenetic tree from which the root-to-tip distance for each isolate is calculated, i.e., the evolutionary change between the time to most recent common ancestor of the sample (TMRCA) and the sampling time of each respective isolate (Tisolate-1, Tisolate-2, Tisolate-n), usually measured as the number of nucleotide substitutions per site over the respective time period. (B) The isolates are plotted based on their date of isolation (x-axis) and root-to-tip divergence, i.e., the number of nucleotide substitutions per site (y-axis). The linear regression provides a measure of temporal signal (R2) as well as molecular clock rate assuming a strict clock is to be applied (slope). In this example, an R2 of 0.432 suggests evidence of a temporal signal in the data, while the (strict) molecular clock rate is estimated at 3.37 × 10−6 subs/site/year.
Jof 07 00661 g001
Jof 07 00661 g002 550
Figure 2. How recombination affects molecular clock-like processes and interpretation of the underlying phylogeny (adapted from Hein et al., 2005). Blue indicates ancestral material, green indicates non-ancestral material. Assuming there is only one recombination event in the history of the sample, to the left of the recombination event, there is one local tree, and to the right, there is another local tree. This can alter the inferred overall phylogeny in three ways: (A) looking back in time, if the two recombining sequences merge with each other before coalescing with any other sequence, the trees will be identical and the recombination event will likely be undetectable; (B) if one of the recombining sequences coalesces with a difference sequence (in this case, S1) before merging with the other recombining sequence, the unrooted tree topology will remain the same, but branch lengths will change; (C) if two or more sequences coalesce with the recombining sequences before the two recombining sequences merge, the tree topology will change.
Figure 2. How recombination affects molecular clock-like processes and interpretation of the underlying phylogeny (adapted from Hein et al., 2005). Blue indicates ancestral material, green indicates non-ancestral material. Assuming there is only one recombination event in the history of the sample, to the left of the recombination event, there is one local tree, and to the right, there is another local tree. This can alter the inferred overall phylogeny in three ways: (A) looking back in time, if the two recombining sequences merge with each other before coalescing with any other sequence, the trees will be identical and the recombination event will likely be undetectable; (B) if one of the recombining sequences coalesces with a difference sequence (in this case, S1) before merging with the other recombining sequence, the unrooted tree topology will remain the same, but branch lengths will change; (C) if two or more sequences coalesce with the recombining sequences before the two recombining sequences merge, the tree topology will change.
Jof 07 00661 g002
Jof 07 00661 g003 550
Figure 3. Examples of tree topology from (A) a population with constant population size and (B) a population undergoing exponential growth. High levels of recombination in a set of isolates can lengthen terminal branches, leading to trees that resemble (B) and resulting in incorrect inference relating to population dynamics over time.
Figure 3. Examples of tree topology from (A) a population with constant population size and (B) a population undergoing exponential growth. High levels of recombination in a set of isolates can lengthen terminal branches, leading to trees that resemble (B) and resulting in incorrect inference relating to population dynamics over time.
Jof 07 00661 g003
Jof 07 00661 g004 550
Figure 4. The fungal temporal analysis roadmap—recommendations for future fungal phylogenetic dating studies to ensure biology is taken into account and phylogenies are valid and replicable.
Figure 4. The fungal temporal analysis roadmap—recommendations for future fungal phylogenetic dating studies to ensure biology is taken into account and phylogenies are valid and replicable.
Jof 07 00661 g004
Table
Table 1. Software tools available for dating phylogenies.
Table 1. Software tools available for dating phylogenies.
TaskTool/AlgorithmDescriptionReference
Identify loci of recombinationGubbinsIteratively identifies loci of recombination and simultaneously constructs
a phylogeny based on point mutations outside of these identified regions.		Croucher et al., Nucleic Acids Res. 2014, doi:10.1093/nar/gku1196
ClonalFrameMLMaximum likelihood inference to
detect loci of recombination and
simultaneously construct a phylogeny accounting for this recombination.		Didelot and Wilson, PLoS Comput. Biol. 2015, doi:10.1371/journal.pcbi.10004041
fastGEARIdentifies population genetic structure of an (species-wide) alignment and detects recombination, both recent and ancestral, between inferred lineages as well as recent recombination from
external origins using a hidden
Markov model.		Mostowy et al., Mol. Biol. Evol. 2017, doi:10.1093/molbev/msx066
Phylogenetic
inference		BEAST/BEAST2Bayesian MCMC algorithm to jointly estimate a phylogeny and its associated parameters (i.e., effective population size, TMRCA, clock rate, etc.)Drummond and Rambaut, BMC Evol. Biol. 2007, doi:10.1017/CBO9781139095112.007
MrBayesBayesian MCMC inference and model choice across a range of phylogenetic and evolutionary models Huelsenbeck and Ronquist, Bioinformatics. 2001, 17 (8), 754–755
IQ-TreeStochastic tree-searching algorithm to identify the highest likelihood tree (output in nucleotide substitutions only, not calendar time)Nguyen et al., Mol. Biol. Evol. 2015, 32(1):268–74.
PhyMLML inference of phylogenetic relationships between divergent populations, utilising subtree pruning and regrafting (SPR) and approximate likelihood-ratio test (aLRT) approaches.Guindon et al., Systematic Biol. 2010,
59 (3), 307–321.
RAxMLML inference of phylogenetic relationships between divergent populations utilising parsimony and heuristic
subtree rearrangements.		Stamatakis, Bioinformatics. 2014,
30 (9), 1312–1313.
PHYLIPPackage of programmes for phylogenetic inference including parsimony, distance matrix and ML methods, bootstrapping and consensus trees.Felsenstein, Cladistics. 1989,
5 (2), 163–166.
SNPhyloPipeline utilising ML to reconstruct phylogenies based on SNP data.Lee et al., BMC Genomics. 2014,
15, 162.
Molecular clock rate/Divergence time estimationTreedaterR package to apply an evolutionary timescale to date and root a phylogeny (i.e., transforms branch lengths from number of nucleotide substitutions to calendar time) and estimate TMRCA. Molecular clock test function tests for appropriate clock model (relaxed
vs. strict).		Volz and Frost, Virus Evol. 2017, doi:10.1093/ve/vex025
PhyTimeA tool in the PhyML package that
estimates divergence dates in
a Bayesian setting.		Guindon, Systematic Biol. 2013,
62 (1), 22–34
Top of Form
Bottom of Form
Phylogeny viewer/editorFigtreeGraphical viewer of phylogenetic trees and to produce publication-ready
figures. Particularly suited to trees produced by BEAST.		http://tree.bio.ed.ac.uk/software/figtree/ accessed on 11 August 2021
IcytreeA simple browser-based phylogenetic tree viewer.https://icytree.org/ accessed on 11 August 2021
GGTREEAn R package for programmable
visualisation and annotation of
phylogenetic trees.		Guangchuang et al., Methods in Ecology and Evolution. 8 (1), 28–36
N.B. This list is by no means exhaustive. Readers can find a more comprehensive database of programs relevant to some of these tasks at http://methodspopgen.com accessed on 11 August 2021.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Edwards, H.M.; Rhodes, J. Accounting for the Biological Complexity of Pathogenic Fungi in Phylogenetic Dating. J. Fungi 2021, 7, 661. https://doi.org/10.3390/jof7080661

AMA Style

Edwards HM, Rhodes J. Accounting for the Biological Complexity of Pathogenic Fungi in Phylogenetic Dating. Journal of Fungi. 2021; 7(8):661. https://doi.org/10.3390/jof7080661

Chicago/Turabian Style

Edwards, Hannah M., and Johanna Rhodes. 2021. "Accounting for the Biological Complexity of Pathogenic Fungi in Phylogenetic Dating" Journal of Fungi 7, no. 8: 661. https://doi.org/10.3390/jof7080661

APA Style

Edwards, H. M., & Rhodes, J. (2021). Accounting for the Biological Complexity of Pathogenic Fungi in Phylogenetic Dating. Journal of Fungi, 7(8), 661. https://doi.org/10.3390/jof7080661

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop