Frequency-dependent selection in vaccine-associated pneumococcal population dynamics

Recent approaches in computational modelling for controlling pathogen threats

In this review, we assess the status of computational modelling of pathogens. We focus on three disparate but interlinked research areas that produce models with very different spatial and temporal scope. First, we examine antimicrobial resistance (AMR). Many mechanisms of AMR are not well understood. As a result, it is hard to measure the current incidence of AMR, predict the future incidence, and design strategies to preserve existing antibiotic effectiveness. Next, we look at how to choose the finite number of bacterial strains that can be included in a vaccine. To do this, we need to understand what happens to vaccine and non-vaccine strains after vaccination programmes. Finally, we look at within-host modelling of antibody dynamics. The SARS-CoV-2 pandemic produced huge amounts of antibody data, prompting improvements in this area of modelling. We finish by discussing the challenges that persist in understanding these complex biological systems.

mSphere of Influence: Predicting the evolution of pathogen populations

Tatum D. Mortimer works in the field of pathogen population genomics and evolution. In this mSphere of Influence article, she reflects on how "Frequency-dependent selection can forecast evolution in Streptococcus pneumoniae" by Azarian et al. and "Contingency, repeatability, and predictability in the evolution of a prokaryotic pangenome" by Beavan et al. made an impact on her by highlighting the ways in which genomic data can be used to predict pathogen evolution.

Geographical migration and fitness dynamics of Streptococcus pneumoniae

Streptococcus pneumoniae is a leading cause of pneumonia and meningitis worldwide. Many different serotypes co-circulate endemically in any one location1,2. The extent and mechanisms of spread and vaccine-driven changes in fitness and antimicrobial resistance remain largely unquantified. Here using geolocated genome sequences from South Africa (n = 6,910, collected from 2000 to 2014), we developed models to reconstruct spread, pairing detailed human mobility data and genomic data. Separately, we estimated the population-level changes in fitness of strains that are included (vaccine type (VT)) and not included (non-vaccine type (NVT)) in pneumococcal conjugate vaccines, first implemented in South Africa in 2009. Differences in strain fitness between those that are and are not resistant to penicillin were also evaluated. We found that pneumococci only become homogenously mixed across South Africa after 50 years of transmission, with the slow spread driven by the focal nature of human mobility. Furthermore, in the years following vaccine implementation, the relative fitness of NVT compared with VT strains increased (relative risk of 1.68; 95% confidence interval of 1.59-1.77), with an increasing proportion of these NVT strains becoming resistant to penicillin. Our findings point to highly entrenched, slow transmission and indicate that initial vaccine-linked decreases in antimicrobial resistance may be transient.

Long-term evolution of Streptococcus mitis and Streptococcus pneumoniae leads to higher genetic diversity within rather than between human populations

Evaluation of the apportionment of genetic diversity of human bacterial commensals within and between human populations is an important step in the characterization of their evolutionary potential. Recent studies showed a correlation between the genomic diversity of human commensal strains and that of their host, but the strength of this correlation and of the geographic structure among human populations is a matter of debate. Here, we studied the genomic diversity and evolution of the phylogenetically related oro-nasopharyngeal healthy-carriage Streptococcus mitis and Streptococcus pneumoniae, whose lifestyles range from stricter commensalism to high pathogenic potential. A total of 119 S. mitis genomes showed higher within- and among-host variation than 810 S. pneumoniae genomes in European, East Asian and African populations. Summary statistics of the site-frequency spectrum for synonymous and non-synonymous variation and ABC modelling showed this difference to be due to higher ancestral bacterial population effective size (Ne) in S. mitis, whose genomic variation has been maintained close to mutation-drift equilibrium across (at least many) generations, whereas S. pneumoniae has been expanding from a smaller ancestral bacterial population. Strikingly, both species show limited differentiation among human populations. As genetic differentiation is inversely proportional to the product of effective population size and migration rate (Nem), we argue that large Ne have led to similar differentiation patterns, even if m is very low for S. mitis. We conclude that more diversity within than among human populations and limited population differentiation must be common features of the human microbiome due to large Ne.

Detecting co-selection through excess linkage disequilibrium in bacterial genomes

Population genomics has revolutionized our ability to study bacterial evolution by enabling data-driven discovery of the genetic architecture of trait variation. Genome-wide association studies (GWAS) have more recently become accompanied by genome-wide epistasis and co-selection (GWES) analysis, which offers a phenotype-free approach to generating hypotheses about selective processes that simultaneously impact multiple loci across the genome. However, existing GWES methods only consider associations between distant pairs of loci within the genome due to the strong impact of linkage-disequilibrium (LD) over short distances. Based on the general functional organisation of genomes it is nevertheless expected that majority of co-selection and epistasis will act within relatively short genomic proximity, on co-variation occurring within genes and their promoter regions, and within operons. Here, we introduce LDWeaver, which enables an exhaustive GWES across both short- and long-range LD, to disentangle likely neutral co-variation from selection. We demonstrate the ability of LDWeaver to efficiently generate hypotheses about co-selection using large genomic surveys of multiple major human bacterial pathogen species and validate several findings using functional annotation and phenotypic measurements. Our approach will facilitate the study of bacterial evolution in the light of rapidly expanding population genomic data.

The balance between intrinsic and ecological fitness defines new regimes in eco-evolutionary population dynamics

Selection upon intrinsic fitness differences is one of the most basic mechanisms of evolution, fundamental to all biology. Equally, within macroscopic populations and microscopic environments, ecological interactions influence evolution. Direct experimental evidence of ecological selection between microscopic agents continues to grow. Whilst eco-evolutionary dynamics describes how interactions influence population fitness and composition, we build a model that allows ecological aspects of these interactions to fall on a spectrum independent of the intrinsic fitness of the population. With our mathematical framework, we show how ecological interactions between mutating populations modify the estimated evolutionary trajectories expected from monoculture fitnesses alone. We derive and validate analytical stationary solutions to our partial differential equations that depend on intrinsic and ecological terms, and mutation rates. We determine cases in which these interactions modify evolution in such ways as to, for example, maintain or invert existing monoculture fitness differences. This work discusses the importance of understanding ecological and intrinsic selection effects to avoid misleading conclusions from experiments and defines new ways to assess this balance from experimental results. Using published experimental data, we also show evidence that real microbiological systems can span intrinsic fitness dominant and ecological-effect dominant regimes and that ecological contributions can change with an environment to exaggerate or counteract the composite populations' intrinsic fitness differences.

Mathematical Modelling of Parasite Dynamics: A Stochastic Simulation-Based Approach and Parameter Estimation via Modified Sequential-Type Approximate Bayesian Computation

The development of mathematical models for studying newly emerging and re-emerging infectious diseases has gained momentum due to global events. The gyrodactylid-fish system, like many host-parasite systems, serves as a valuable resource for ecological, evolutionary, and epidemiological investigations owing to its ease of experimental manipulation and long-term monitoring. Although this system has an existing individual-based model, it falls short in capturing information about species-specific microhabitat preferences and other biological details for different Gyrodactylus strains across diverse fish populations. This current study introduces a new individual-based stochastic simulation model that uses a hybrid τ -leaping algorithm to incorporate this essential data, enhancing our understanding of the complexity of the gyrodactylid-fish system. We compare the infection dynamics of three gyrodactylid strains across three host populations. A modified sequential-type approximate Bayesian computation (ABC) method, based on sequential Monte Carlo and sequential importance sampling, is developed. Additionally, we establish two penalised local-linear regression methods (based on L1 and L2 regularisations) for ABC post-processing analysis to fit our model using existing empirical data. With the support of experimental data and the fitted mathematical model, we address open biological questions for the first time and propose directions for future studies on the gyrodactylid-fish system. The adaptability of the mathematical model extends beyond the gyrodactylid-fish system to other host-parasite systems. Furthermore, the modified ABC methodologies provide efficient calibration for other multi-parameter models characterised by a large set of correlated or independent summary statistics.

Convergent impact of vaccination and antibiotic pressures on pneumococcal populations

Streptococcus pneumoniae is a remarkably adaptable and successful human pathogen, playing dual roles of both asymptomatic carriage in the nasopharynx and invasive disease including pneumonia, bacteremia, and meningitis. Efficacious vaccines and effective antibiotic therapies are critical to mitigating morbidity and mortality. However, clinical interventions can be rapidly circumvented by the pneumococcus by its inherent proclivity for genetic exchange. This leads to an underappreciated interplay between vaccine and antibiotic pressures on pneumococcal populations. Circulating populations have undergone dramatic shifts due to the introduction of capsule-based vaccines of increasing valency imparting strong selective pressures. These alterations in population structure have concurrent consequences on the frequency of antibiotic resistance profiles in the population. This review will discuss the interactions of these two selective forces. Understanding and forecasting the drivers of antibiotic resistance and capsule switching are of critical importance for public health, particularly for such a genetically promiscuous pathogen as S. pneumoniae.

Pneumococcal population genomics changes during the early time period of conjugate vaccine uptake in southern India

Streptococcus pneumoniae is a major cause of invasive disease of young children in low- and middle-income countries. In southern India, pneumococcal conjugate vaccines (PCVs) that can prevent invasive pneumococcal disease began to be used more frequently after 2015. To characterize pneumococcal evolution during the early time period of PCV uptake in southern India, genomes were sequenced and selected characteristics were determined for 402 invasive isolates collected from children <5 years of age during routine surveillance from 1991 to 2020. Overall, the prevalence and diversity of vaccine type (VT) and non-vaccine type (NVT) isolates did not significantly change post-uptake of PCV. Individually, serotype 1 and global pneumococcal sequence cluster (GPSC or strain lineage) 2 significantly decreased, whereas serotypes 6B, 9V and 19A and GPSCs 1, 6, 10 and 23 significantly increased in proportion post-uptake of PCV. Resistance determinants to penicillin, erythromycin, co-trimoxazole, fluoroquinolones and tetracycline, and multidrug resistance significantly increased in proportion post-uptake of PCV and especially among VT isolates. Co-trimoxazole resistance determinants were common pre- and post-uptake of PCV (85 and 93 %, respectively) and experienced the highest rates of recombination in the genome. Accessory gene frequencies were seen to be changing by small amounts across the frequency spectrum specifically among VT isolates, with the largest changes linked to antimicrobial resistance determinants. In summary, these results indicate that as of 2020 this pneumococcal population was not yet approaching a PCV-induced equilibrium and they highlight changes related to antimicrobial resistance. Augmenting PCV coverage and prudent use of antimicrobials are needed to counter invasive pneumococcal disease in this region.

Designing optimal behavioral experiments using machine learning

Computational models are powerful tools for understanding human cognition and behavior. They let us express our theories clearly and precisely and offer predictions that can be subtle and often counter-intuitive. However, this same richness and ability to surprise means our scientific intuitions and traditional tools are ill-suited to designing experiments to test and compare these models. To avoid these pitfalls and realize the full potential of computational modeling, we require tools to design experiments that provide clear answers about what models explain human behavior and the auxiliary assumptions those models must make. Bayesian optimal experimental design (BOED) formalizes the search for optimal experimental designs by identifying experiments that are expected to yield informative data. In this work, we provide a tutorial on leveraging recent advances in BOED and machine learning to find optimal experiments for any kind of model that we can simulate data from, and show how by-products of this procedure allow for quick and straightforward evaluation of models and their parameters against real experimental data. As a case study, we consider theories of how people balance exploration and exploitation in multi-armed bandit decision-making tasks. We validate the presented approach using simulations and a real-world experiment. As compared to experimental designs commonly used in the literature, we show that our optimal designs more efficiently determine which of a set of models best account for individual human behavior, and more efficiently characterize behavior given a preferred model. At the same time, formalizing a scientific question such that it can be adequately addressed with BOED can be challenging and we discuss several potential caveats and pitfalls that practitioners should be aware of. We provide code to replicate all analyses as well as tutorial notebooks and pointers to adapt the methodology to different experimental settings.

Genomic and panproteomic analysis of the development of infant immune responses to antigenically-diverse pneumococci

Streptococcus pneumoniae (pneumococcus) is a nasopharyngeal commensal and respiratory pathogen. This study characterises the immunoglobulin G (IgG) repertoire recognising pneumococci from birth to 24 months old (mo) in a prospectively-sampled cohort of 63 children using a panproteome array. IgG levels are highest at birth, due to transplacental transmission of maternal antibodies. The subsequent emergence of responses to individual antigens exhibit distinct kinetics across the cohort. Stable differences in the strength of individuals' responses, correlating with maternal IgG concentrations, are established by 6 mo. By 12 mo, children develop unique antibody profiles that are boosted by re-exposure. However, some proteins only stimulate substantial responses in adults. Integrating genomic data on nasopharyngeal colonisation demonstrates rare pneumococcal antigens can elicit strong IgG levels post-exposure. Quantifying such responses to the diverse core loci (DCL) proteins is complicated by cross-immunity between variants. In particular, the conserved N terminus of DCL protein zinc metalloprotease B provokes the strongest early IgG responses. DCL proteins' ability to inhibit mucosal immunity likely explains continued pneumococcal carriage despite hosts' polyvalent antibody repertoire. Yet higher IgG levels are associated with reduced incidence, and severity, of pneumonia, demonstrating the importance of the heterogeneity in response strength and kinetics across antigens and individuals.

Microevolution and Its Impact on Hypervirulence, Antimicrobial Resistance, and Vaccine Escape in Neisseria meningitidis

Neisseria meningitidis is commensal of the human pharynx and occasionally invades the host, causing the life-threatening illness invasive meningococcal disease. The meningococcus is a highly diverse and adaptable organism thanks to natural competence, a propensity for recombination, and a highly repetitive genome. These mechanisms together result in a high level of antigenic variation to invade diverse human hosts and evade their innate and adaptive immune responses. This review explores the ways in which this diversity contributes to the evolutionary history and population structure of the meningococcus, with a particular focus on microevolution. It examines studies on meningococcal microevolution in the context of within-host evolution and persistent carriage; microevolution in the context of meningococcal outbreaks and epidemics; and the potential of microevolution to contribute to antimicrobial resistance and vaccine escape. A persistent theme is the idea that the process of microevolution contributes to the development of new hyperinvasive meningococcal variants. As such, microevolution in this species has significant potential to drive future public health threats in the form of hypervirulent, antibiotic-resistant, vaccine-escape variants. The implications of this on current vaccination strategies are explored.

Prediction of post-PCV13 pneumococcal evolution using invasive disease data enhanced by inverse-invasiveness weighting

Background

After introduction of pneumococcal conjugate vaccines (PCVs), serotype replacement occurred in the population of Streptococcus pneumoniae. Predicting which pneumococcal clones and serotypes will become more common in carriage after vaccination can enhance vaccine design and public health interventions, while also improving our understanding of pneumococcal evolution. We sought to use invasive disease data to assess how well negative frequency-dependent selection (NFDS) models could explain pneumococcal carriage population evolution in the post-PCV13 epoch by weighting invasive data to approximate strain proportions in the carriage population.

Methods

Invasive pneumococcal isolates were collected and sequenced during 1998-2018 by the Active Bacterial Core surveillance (ABCs) from the Centers for Disease Control and Prevention (CDC). To predict the post-PCV13 population dynamics in the carriage population using a NFDS model, all genomic data were processed under a bioinformatic pipeline of assembly, annotation, and pangenome analysis to define genetically similar sequence clusters (i.e., strains) and a set of accessory genes present in 5% to 95% of the isolates. The NFDS model predicted the strain proportion by calculating the post-vaccine strain composition in the weighted invasive disease population that would best match pre-vaccine accessory gene frequencies. To overcome the biases of invasive disease data, serotype-specific inverse-invasiveness weights were defined as the ratio of the proportion of the serotype in the carriage data to the proportion in the invasive data, using data from 1998-2001 in the United States, before conjugate vaccine introduction. The weights were applied to adjust both the observed strain proportion and the accessory gene frequencies.

Results

Inverse-invasiveness weighting increased the correlation of accessory gene frequencies between invasive and carriage data with reduced residuals in linear or logit scale for pre-vaccine, post-PCV7, and post-PCV13. Similarly, weighting increased the correlation of accessory gene frequencies between different time periods in the invasive data. By weighting the invasive data, we were able to use the NFDS model to predict strain proportions in the carriage population in the post-PCV13 epoch, with the adjusted R-squared between predicted and observed strain proportions increasing from 0.176 to 0.544 after weighting.

Conclusions

The weighting system adjusted the invasive disease surveillance data to better represent the carriage population of S. pneumoniae. The NFDS mechanism predicted the strain proportions in the projected carriage population as estimated from the weighted invasive disease frequencies in the post-PCV13 epoch. Our methods enrich the value of genomic sequences from invasive disease surveillance, which is readily available, easy to collect, and of direct interest to public health.

Pneumococcal population dynamics: Investigating vaccine-induced changes through multiscale modelling

The mechanisms behind vaccine-induced strain replacement in the pneumococcus remain poorly understood. There is emerging evidence that distinct pneumococcal lineages can co-colonise for significant time periods, and that novel recombinants can readily emerge during natural colonisation. Despite this, patterns of post-vaccine replacement are indicative of competition between specific lineages. Here, we develop a multiscale transmission model to investigate explicitly how within host dynamics shape observed ecological patterns, both pre- and post-vaccination. Our model framework explores competition between and within strains defined by distinct antigenic, metabolic and resistance profiles. We allow for strains to freely co-colonise and recombine within hosts, and consider how each of these types may contribute to a strain's overall fitness. Our results suggest that antigenic and resistance profiles are key drivers of post-vaccine success.

The metabolic, virulence and antimicrobial resistance profiles of colonising Streptococcus pneumoniae shift after PCV13 introduction in urban Malawi

Streptococcus pneumoniae causes substantial mortality among children under 5-years-old worldwide. Polysaccharide conjugate vaccines (PCVs) are highly effective at reducing vaccine serotype disease, but emergence of non-vaccine serotypes and persistent nasopharyngeal carriage threaten this success. We investigated the hypothesis that following vaccine, adapted pneumococcal genotypes emerge with the potential for vaccine escape. We genome sequenced 2804 penumococcal isolates, collected 4-8 years after introduction of PCV13 in Blantyre, Malawi. We developed a pipeline to cluster the pneumococcal population based on metabolic core genes into "Metabolic genotypes" (MTs). We show that S. pneumoniae population genetics are characterised by emergence of MTs with distinct virulence and antimicrobial resistance (AMR) profiles. Preliminary in vitro and murine experiments revealed that representative isolates from emerging MTs differed in growth, haemolytic, epithelial infection, and murine colonisation characteristics. Our results suggest that in the context of PCV13 introduction, pneumococcal population dynamics had shifted, a phenomenon that could further undermine vaccine control and promote spread of AMR.

The influence of inter-regional delays in generating large-scale brain networks of phase synchronization

Large-scale networks of phase synchronization are considered to regulate the communication between brain regions fundamental to cognitive function, but the mapping to their structural substrates, i.e., the structure-function relationship, remains poorly understood. Biophysical Network Models (BNMs) have demonstrated the influences of local oscillatory activity and inter-regional anatomical connections in generating alpha-band (8-12 Hz) networks of phase synchronization observed with Electroencephalography (EEG) and Magnetoencephalography (MEG). Yet, the influence of inter-regional conduction delays remains unknown. In this study, we compared a BNM with standard "distance-dependent delays", which assumes constant conduction velocity, to BNMs with delays specified by two alternative methods accounting for spatially varying conduction velocities, "isochronous delays" and "mixed delays". We followed the Approximate Bayesian Computation (ABC) workflow, i) specifying neurophysiologically informed prior distributions of BNM parameters, ii) verifying the suitability of the prior distributions with Prior Predictive Checks, iii) fitting each of the three BNMs to alpha-band MEG resting-state data (N = 75) with Bayesian optimization for Likelihood-Free Inference (BOLFI), and iv) choosing between the fitted BNMs with ABC model comparison on a separate MEG dataset (N = 30). Prior Predictive Checks revealed the range of dynamics generated by each of the BNMs to encompass those seen in the MEG data, suggesting the suitability of the prior distributions. Fitting the models to MEG data yielded reliable posterior distributions of the parameters of each of the BNMs. Finally, model comparison revealed the BNM with "distance-dependent delays", as the most probable to describe the generation of alpha-band networks of phase synchronization seen in MEG. These findings suggest that distance-dependent delays might contribute to the neocortical architecture of human alpha-band networks of phase synchronization. Hence, our study illuminates the role of inter-regional delays in generating the large-scale networks of phase synchronization that might subserve the communication between regions vital to cognition.

Accurate and fast graph-based pangenome annotation and clustering with ggCaller

Bacterial genomes differ in both gene content and sequence mutations, which underlie extensive phenotypic diversity, including variation in susceptibility to antimicrobials or vaccine-induced immunity. To identify and quantify important variants, all genes within a population must be predicted, functionally annotated, and clustered, representing the "pangenome." Despite the volume of genome data available, gene prediction and annotation are currently conducted in isolation on individual genomes, which is computationally inefficient and frequently inconsistent across genomes. Here, we introduce the open-source software graph-gene-caller (ggCaller). ggCaller combines gene prediction, functional annotation, and clustering into a single workflow using population-wide de Bruijn graphs, removing redundancy in gene annotation and resulting in more accurate gene predictions and orthologue clustering. We applied ggCaller to simulated and real-world bacterial data sets containing hundreds or thousands of genomes, comparing it to current state-of-the-art tools. ggCaller has considerable speed-ups with equivalent or greater accuracy, particularly with data sets containing complex sources of error, such as assembly contamination or fragmentation. ggCaller is also an important extension to bacterial genome-wide association studies, enabling querying of annotated graphs for functional analyses. We highlight this application by functionally annotating DNA sequences with significant associations to tetracycline and macrolide resistance in Streptococcus pneumoniae, identifying key resistance determinants that were missed when using only a single reference genome. ggCaller is a novel bacterial genome analysis tool with applications in bacterial evolution and epidemiology.

Geographic migration and vaccine-induced fitness changes of Streptococcus pneumoniae

Streptococcus pneumoniae is a leading cause of pneumonia and meningitis worldwide. Many different serotypes co-circulate endemically in any one location. The extent and mechanisms of spread, and vaccine-driven changes in fitness and antimicrobial resistance (AMR), remain largely unquantified. Using geolocated genome sequences from South Africa (N=6910, 2000-2014) we developed models to reconstruct spread, pairing detailed human mobility data and genomic data. Separately we estimated the population level changes in fitness of strains that are (vaccine type, VT) and are not (non-vaccine type, NVT) included in the vaccine, first implemented in 2009, as well as differences in strain fitness between those that are and are not resistant to penicillin. We estimated that pneumococci only become homogenously mixed across South Africa after about 50 years of transmission, with the slow spread driven by the focal nature of human mobility. Further, in the years following vaccine implementation the relative fitness of NVT compared to VT strains increased (RR: 1.29 [95% CI 1.20-1.37]) - with an increasing proportion of these NVT strains becoming penicillin resistant. Our findings point to highly entrenched, slow transmission and indicate that initial vaccine-linked decreases in AMR may be transient.

Post-vaccine epidemiology of serotype 3 pneumococci identifies transformation inhibition through prophage-driven alteration of a non-coding RNA

BACKGROUND

The respiratory pathogen Streptococcus pneumoniae (the pneumococcus) is a genetically diverse bacterium associated with over 101 immunologically distinct polysaccharide capsules (serotypes). Polysaccharide conjugate vaccines (PCVs) have successfully eliminated multiple targeted serotypes, yet the mucoid serotype 3 has persisted despite its inclusion in PCV13. This capsule type is predominantly associated with a single globally disseminated strain, GPSC12 (clonal complex 180).

METHODS

A genomic epidemiology study combined previous surveillance datasets of serotype 3 pneumococci to analyse the population structure, dynamics, and differences in rates of diversification within GPSC12 during the period of PCV introductions. Transcriptomic analyses, whole genome sequencing, mutagenesis, and electron microscopy were used to characterise the phenotypic impact of loci hypothesised to affect this strain's evolution.

RESULTS

GPSC12 was split into clades by a genomic analysis. Clade I, the most common, rarely underwent transformation, but was typically infected with the prophage ϕOXC141. Prior to the introduction of PCV13, this clade's composition shifted towards a ϕOXC141-negative subpopulation in a systematically sampled UK collection. In the post-PCV13 era, more rapidly recombining non-Clade I isolates, also ϕOXC141-negative, have risen in prevalence. The low in vitro transformation efficiency of a Clade I isolate could not be fully explained by the ~100-fold reduction attributable to the serotype 3 capsule. Accordingly, prophage ϕOXC141 was found to modify csRNA3, a non-coding RNA that inhibits the induction of transformation. This alteration was identified in ~30% of all pneumococci and was particularly common in the unusually clonal serotype 1 GPSC2 strain. RNA-seq and quantitative reverse transcriptase PCR experiments using a genetically tractable pneumococcus demonstrated the altered csRNA3 was more effective at inhibiting production of the competence-stimulating peptide pheromone. This resulted in a reduction in the induction of competence for transformation.

CONCLUSION

This interference with the quorum sensing needed to induce competence reduces the risk of the prophage being deleted by homologous recombination. Hence the selfish prophage-driven alteration of a regulatory RNA limits cell-cell communication and horizontal gene transfer, complicating the interpretation of post-vaccine population dynamics.

Genomics and pathotypes of the many faces of Escherichia coli

Escherichia coli is the most researched microbial organism in the world. Its varied impact on human health, consisting of commensalism, gastrointestinal disease, or extraintestinal pathologies, has generated a separation of the species into at least eleven pathotypes (also known as pathovars). These are broadly split into two groups, intestinal pathogenic E. coli (InPEC) and extraintestinal pathogenic E. coli (ExPEC). However, components of E. coli's infinite open accessory genome are horizontally transferred with substantial frequency, creating pathogenic hybrid strains that defy a clear pathotype designation. Here, we take a birds-eye view of the E. coli species, characterizing it from historical, clinical, and genetic perspectives. We examine the wide spectrum of human disease caused by E. coli, the genome content of the bacterium, and its propensity to acquire, exchange, and maintain antibiotic resistance genes and virulence traits. Our portrayal of the species also discusses elements that have shaped its overall population structure and summarizes the current state of vaccine development targeted at the most frequent E. coli pathovars. In our conclusions, we advocate streamlining efforts for clinical reporting of ExPEC, and emphasize the pathogenic potential that exists throughout the entire species.

Prokaryotic Pangenomes Act as Evolving Ecosystems

Understanding adaptation to the local environment is a central tenet and a major focus of evolutionary biology. But this is only part of the adaptionist story. In addition to the external environment, one of the main drivers of genome composition is genetic background. In this perspective, I argue that there is a growing body of evidence that intra-genomic selective pressures play a significant part in the composition of prokaryotic genomes and play a significant role in the origin, maintenance and structuring of prokaryotic pangenomes.

The potential of genomics for infectious disease forecasting

Genomic technologies have led to tremendous gains in understanding how pathogens function, evolve and interact. Pathogen diversity is now measurable at high precision and resolution, in part because over the past decade, sequencing technologies have increased in speed and capacity, at decreased cost. Alongside this, the use of models that can forecast emergence and size of infectious disease outbreaks has risen, highlighted by the coronavirus disease 2019 pandemic but also due to modelling advances that allow for rapid estimates in emerging outbreaks to inform monitoring, coordination and resource deployment. However, genomics studies have remained largely retrospective. While they contain high-resolution views of pathogen diversification and evolution in the context of selection, they are often not aligned with designing interventions. This is a missed opportunity because pathogen diversification is at the core of the most pressing infectious public health challenges, and interventions need to take the mechanisms of virulence and understanding of pathogen diversification into account. In this Perspective, we assess these converging fields, discuss current challenges facing both surveillance specialists and modellers who want to harness genomic data, and propose next steps for integrating longitudinally sampled genomic data with statistical learning and interpretable modelling to make reliable predictions into the future.

Mandrake: visualizing microbial population structure by embedding millions of genomes into a low-dimensional representation

In less than a decade, population genomics of microbes has progressed from the effort of sequencing dozens of strains to thousands, or even tens of thousands of strains in a single study. There are now hundreds of thousands of genomes available even for a single bacterial species, and the number of genomes is expected to continue to increase at an accelerated pace given the advances in sequencing technology and widespread genomic surveillance initiatives. This explosion of data calls for innovative methods to enable rapid exploration of the structure of a population based on different data modalities, such as multiple sequence alignments, assemblies and estimates of gene content across different genomes. Here, we present Mandrake, an efficient implementation of a dimensional reduction method tailored for the needs of large-scale population genomics. Mandrake is capable of visualizing population structure from millions of whole genomes, and we illustrate its usefulness with several datasets representing major pathogens. Our method is freely available both as an analysis pipeline (https://github.com/johnlees/mandrake) and as a browser-based interactive application (https://gtonkinhill.github.io/mandrake-web/).

This article is part of a discussion meeting issue ‘Genomic population structures of microbial pathogens’.

Emergence of a multidrug-resistant and virulent Streptococcus pneumoniae lineage mediates serotype replacement after PCV13: an international whole-genome sequencing study

BACKGROUND

Serotype 24F is one of the emerging pneumococcal serotypes after the introduction of pneumococcal conjugate vaccine (PCV). We aimed to identify lineages driving the increase of serotype 24F in France and place these findings into a global context.

METHODS

Whole-genome sequencing was performed on a collection of serotype 24F pneumococci from asymptomatic colonisation (n=229) and invasive disease (n=190) isolates among individuals younger than 18 years in France, from 2003 to 2018. To provide a global context, we included an additional collection of 24F isolates in the Global Pneumococcal Sequencing (GPS) project database for analysis. A Global Pneumococcal Sequence Cluster (GPSC) and a clonal complex (CC) were assigned to each genome. Phylogenetic, evolutionary, and spatiotemporal analysis were conducted using the same 24F collection and supplemented with a global collection of genomes belonging to the lineage of interest from the GPS project database (n=25 590).

FINDINGS

Serotype 24F was identified in numerous countries mainly due to the clonal spread of three lineages: GPSC10 (CC230), GPSC16 (CC156), and GPSC206 (CC7701). GPSC10 was the only multidrug-resistant lineage. GPSC10 drove the increase in 24F in France and had high invasive disease potential. The international dataset of GPSC10 (n=888) revealed that this lineage expressed 16 other serotypes, with only six included in 13-valent PCV (PCV13). All serotype 24F isolates were clustered in a single clade within the GPSC10 phylogeny and long-range transmissions were detected from Europe to other continents. Spatiotemporal analysis showed GPSC10-24F took 3-5 years to spread across France and a rapid change of serotype composition from PCV13 serotype 19A to 24F during the introduction of PCV13 was observed in neighbouring country Spain.

INTERPRETATION

Our work reveals that GPSC10 alone is a challenge for serotype-based vaccine strategy. More systematic investigation to identify lineages like GPSC10 will better inform and improve next-generation preventive strategies against pneumococcal diseases.

FUNDING

Bill & Melinda Gates Foundation, Wellcome Sanger Institute, and the US Centers for Disease Control and Prevention.

Pneumococcal genetic variability in age-dependent bacterial carriage

The characteristics of pneumococcal carriage vary between infants and adults. Host immune factors have been shown to contribute to these age-specific differences, but the role of pathogen sequence variation is currently less well-known. Identification of age-associated pathogen genetic factors could leadto improved vaccine formulations. We therefore performed genome sequencing in a large carriage cohort of children and adults and combined this with data from an existing age-stratified carriage study. We compiled a dictionary of pathogen genetic variation, including serotype, strain, sequence elements, single-nucleotide polymorphisms (SNPs), and clusters of orthologous genes (COGs) for each cohort - all of which were used in a genome-wide association with host age. Age-dependent colonization showed weak evidence of being heritable in the first cohort (h2 = 0.10, 95% CI 0.00-0.69) and stronger evidence in the second cohort (h2 = 0.56, 95% CI 0.23-0.87). We found that serotypes and genetic background (strain) explained a proportion of the heritability in the first cohort (h2serotype = 0.07, 95% CI 0.04-0.14 and h2GPSC = 0.06, 95% CI 0.03-0.13) and the second cohort (h2serotype = 0.11, 95% CI 0.05-0.21 and h2GPSC = 0.20, 95% CI 0.12-0.31). In a meta-analysis of these cohorts, we found one candidate association (p=1.2 × 10-9) upstream of an accessory Sec-dependent serine-rich glycoprotein adhesin. Overall, while we did find a small effect of pathogen genome variation on pneumococcal carriage between child and adult hosts, this was variable between populations and does not appear to be caused by strong effects of individual genes. This supports proposals for adaptive future vaccination strategies that are primarily targeted at dominant circulating serotypes and tailored to the composition of the pathogen populations.

Distinct Streptococcus pneumoniae cause invasive disease in Papua New Guinea

Streptococcus pneumoniae is a key contributor to childhood morbidity and mortality in Papua New Guinea (PNG). For the first time, whole genome sequencing of 174 isolates has enabled detailed characterisation of diverse S. pneumoniae causing invasive disease in young children in PNG, 1989-2014. This study captures the baseline S. pneumoniae population prior to the introduction of 13-valent pneumococcal conjugate vaccine (PCV13) into the national childhood immunisation programme in 2014. Relationships amongst lineages, serotypes and antimicrobial resistance traits were characterised, and the population was viewed in the context of a global collection of isolates. The analyses highlighted adiverse S. pneumoniae population associated with invasive disease in PNG, with 45 unique Global Pneumococcal Sequence Clusters (GPSCs) observed amongst the 174 isolates reflecting multiple lineages observed in PNG that have not been identified in other geographic locations. The majority of isolates were from children with meningitis, of which 52% (n=72) expressed non-PCV13 serotypes. Over a third of isolates were predicted to be resistant to at least one antimicrobial. PCV13 serotype isolates had 10.1 times the odds of being multidrug-resistant (MDR) compared to non-vaccine serotype isolates, and no isolates with GPSCs unique to PNG were MDR. Serotype 2 was the most commonly identified serotype; we identified a highly clonal cluster of serotype 2 isolates unique to PNG, and a distinct second cluster indicative of long-distance transmission. Ongoing surveillance, including whole-genome sequencing, is needed to ascertain the impact of the national PCV13 programme upon the S. pneumoniae population, including serotype replacement and antimicrobial resistance traits.

Genetic background of Cambodian pneumococcal carriage isolates following pneumococcal conjugate vaccine 13

Streptococcus pneumoniae (the pneumococcus) is a leading cause of childhood mortality globally and in Cambodia. It is commensal in the human nasopharynx, occasionally resulting in invasive disease. Monitoring population genetic shifts, characterized by lineage and serotype expansions, as well as antimicrobial-resistance (AMR) patterns is crucial for assessing and predicting the impact of vaccination campaigns. We sought to elucidate the genetic background (global pneumococcal sequence clusters; GPSCs) of pneumococci carried by Cambodian children following perturbation by pneumococcal conjugate vaccine (PCV) 13. We sequenced pre-PCV13 (01/2013-12/2015, N=258) and post-PCV13 carriage isolates (01/2016-02/2017, N=428) and used PopPUNK and SeroBA to determine lineage prevalence and serotype composition. Following PCV13 implementation in Cambodia, we saw expansions of non-vaccine type (NVT) serotypes 23A (GPSC626), 34 (GPSC45) and 6D (GPSC16). We predicted antimicrobial susceptibility using the CDC-AMR pipeline and determined concordance with phenotypic data. The CDC-AMR pipeline had >90 % concordance with the phenotypic antimicrobial-susceptibility testing. We detected a high prevalence of AMR in both expanding non-vaccine serotypes and residual vaccine serotype 6B. Persistently high levels of AMR, specifically persisting multidrug-resistant lineages, warrant concern. The implementation of PCV13 in Cambodia has resulted in NVT serotype expansion reflected in the carriage population and driven by specific genetic backgrounds. Continued monitoring of these GPSCs during the ongoing collection of additional carriage isolates in this population is necessary.

Global spatial dynamics and vaccine-induced fitness changes of Bordetella pertussis

As with other pathogens, competitive interactions between Bordetella pertussis strains drive infection risk. Vaccines are thought to perturb strain diversity through shifts in immune pressures; however, this has rarely been measured because of inadequate data and analytical tools. We used 3344 sequences from 23 countries to show that, on average, there are 28.1 transmission chains circulating within a subnational region, with the number of chains strongly associated with host population size. It took 5 to 10 years for B. pertussis to be homogeneously distributed throughout Europe, with the same time frame required for the United States. Increased fitness of pertactin-deficient strains after implementation of acellular vaccines, but reduced fitness otherwise, can explain long-term genotype dynamics. These findings highlight the role of vaccine policy in shifting local diversity of a pathogen that is responsible for 160,000 deaths annually.

Analysing pneumococcal invasiveness using Bayesian models of pathogen progression rates

The disease burden attributable to opportunistic pathogens depends on their prevalence in asymptomatic colonisation and the rate at which they progress to cause symptomatic disease. Increases in infections caused by commensals can result from the emergence of “hyperinvasive” strains. Such pathogens can be identified through quantifying progression rates using matched samples of typed microbes from disease cases and healthy carriers. This study describes Bayesian models for analysing such datasets, implemented in an RStan package (https://github.com/nickjcroucher/progressionEstimation). The models converged on stable fits that accurately reproduced observations from meta-analyses of Streptococcus pneumoniae datasets. The estimates of invasiveness, the progression rate from carriage to invasive disease, in cases per carrier per year correlated strongly with the dimensionless values from meta-analysis of odds ratios when sample sizes were large. At smaller sample sizes, the Bayesian models produced more informative estimates. This identified historically rare but high-risk S. pneumoniae serotypes that could be problematic following vaccine-associated disruption of the bacterial population. The package allows for hypothesis testing through model comparisons with Bayes factors. Application to datasets in which strain and serotype information were available for S. pneumoniae found significant evidence for within-strain and within-serotype variation in invasiveness. The heterogeneous geographical distribution of these genotypes is therefore likely to contribute to differences in the impact of vaccination in between locations. Hence genomic surveillance of opportunistic pathogens is crucial for quantifying the effectiveness of public health interventions, and enabling ongoing meta-analyses that can identify new, highly invasive variants.

Pneumococcal within-host diversity during colonization, transmission and treatment

Characterizing the genetic diversity of pathogens within the host promises to greatly improve surveillance and reconstruction of transmission chains. For bacteria, it also informs our understanding of inter-strain competition and how this shapes the distribution of resistant and sensitive bacteria. Here we study the genetic diversity of Streptococcus pneumoniae within 468 infants and 145 of their mothers by deep sequencing whole pneumococcal populations from 3,761 longitudinal nasopharyngeal samples. We demonstrate that deep sequencing has unsurpassed sensitivity for detecting multiple colonization, doubling the rate at which highly invasive serotype 1 bacteria were detected in carriage compared with gold-standard methods. The greater resolution identified an elevated rate of transmission from mothers to their children in the first year of the child's life. Comprehensive treatment data demonstrated that infants were at an elevated risk of both the acquisition and persistent colonization of a multidrug-resistant bacterium following antimicrobial treatment. Some alleles were enriched after antimicrobial treatment, suggesting that they aided persistence, but generally purifying selection dominated within-host evolution. Rates of co-colonization imply that in the absence of treatment, susceptible lineages outcompeted resistant lineages within the host. These results demonstrate the many benefits of deep sequencing for the genomic surveillance of bacterial pathogens.

Pneumococcal capsule blocks protection by immunization with conserved surface proteins

Vaccines targeting Streptococcus pneumoniae (Spn) are limited by dependence on capsular polysaccharide and its serotype diversity. More broadly-based approaches using common protein antigens have not resulted in a licensed vaccine. Herein, we used an unbiased, genome-wide approach to find novel vaccine antigens to disrupt carriage modeled in mice. A Tn-Seq screen identified 198 genes required for colonization of which 16 are known to express conserved, immunogenic surface proteins. After testing defined mutants for impaired colonization of infant and adult mice, 5 validated candidates (StkP, PenA/Pbp2a, PgdA, HtrA, and LytD/Pce/CbpE) were used as immunogens. Despite induction of antibody recognizing the Spn cell surface, there was no protection against Spn colonization. There was, however, protection against an unencapsulated Spn mutant. This result correlated with increased antibody binding to the bacterial surface in the absence of capsule. Our findings demonstrate how the pneumococcal capsule interferes with mucosal protection by antibody to common protein targets.

Exact simulation of coupled Wright–Fisher diffusions

In this paper an exact rejection algorithm for simulating paths of the coupled Wright–Fisher diffusion is introduced. The coupled Wright–Fisher diffusion is a family of multivariate Wright–Fisher diffusions that have drifts depending on each other through a coupling term and that find applications in the study of networks of interacting genes. The proposed rejection algorithm uses independent neutral Wright–Fisher diffusions as candidate proposals, which are only needed at a finite number of points. Once a candidate is accepted, the remainder of the path can be recovered by sampling from neutral multivariate Wright–Fisher bridges, for which an exact sampling strategy is also provided. Finally, the algorithm’s complexity is derived and its performance demonstrated in a simulation study.

Phylogeography and resistome of pneumococcal meningitis in West Africa before and after vaccine introduction

Despite contributing to the large disease burden in West Africa, little is known about the genomic epidemiology of Streptococcus pneumoniae which cause meningitis among children under 5 years old in the region. We analysed whole-genome sequencing data from 185 S. pneumoniae isolates recovered from suspected paediatric meningitis cases as part of the World Health Organization (WHO) invasive bacterial diseases surveillance from 2010 to 2016. The phylogeny was reconstructed, accessory genome similarity was computed and antimicrobial-resistance patterns were inferred from the genome data and compared to phenotypic resistance from disc diffusion. We studied the changes in the distribution of serotypes pre- and post-pneumococcal conjugate vaccine (PCV) introduction in the Central and Western sub-regions separately. The overall distribution of non-vaccine, PCV7 (4, 6B, 9V, 14, 18C, 19F and 23F) and additional PCV13 serotypes (1, 3, 5, 6A, 19A and 7F) did not change significantly before and after PCV introduction in the Central region (Fisher's test P value 0.27) despite an increase in the proportion of non-vaccine serotypes to 40 % (n=6) in the post-PCV introduction period compared to 21.9 % (n=14). In the Western sub-region, PCV13 serotypes were more dominant among isolates from The Gambia following the introduction of PCV7, 81 % (n=17), compared to the pre-PCV period in neighbouring Senegal, 51 % (n=27). The phylogeny illustrated the diversity of strains associated with paediatric meningitis in West Africa and highlighted the existence of phylogeographical clustering, with isolates from the same sub-region clustering and sharing similar accessory genome content. Antibiotic-resistance genotypes known to confer resistance to penicillin, chloramphenicol, co-trimoxazole and tetracycline were detected across all sub-regions. However, there was no discernible trend linking the presence of resistance genotypes with the vaccine introduction period or whether the strain was a vaccine or non-vaccine serotype. Resistance genotypes appeared to be conserved within selected sub-clades of the phylogenetic tree, suggesting clonal inheritance. Our data underscore the need for continued surveillance on the emergence of non-vaccine serotypes as well as chloramphenicol and penicillin resistance, as these antibiotics are likely still being used for empirical treatment in low-resource settings. This article contains data hosted by Microreact.

Serotype Patterns of Pneumococcal Disease in Adults Are Correlated With Carriage Patterns in Older Children

BACKGROUND

The importance of specific serotypes causing invasive pneumococcal disease (IPD) differs by age. Data on pneumococcal carriage in different age groups, along with data on serotype-specific invasiveness, could help explain these age-related patterns and their implications for vaccination.

METHODS

Using pneumococcal carriage and disease data from Israel, we evaluated the association between serotype-specific IPD in adults and serotype-specific carriage prevalence among children in different age categories, while adjusting for serotype-specific invasiveness. We estimated carriage prevalence using different age groupings that were selected a priori. The Deviance Information Criterion was used to determine which age groupings of carriage data best fit the adult IPD data. Serotype-specific disease patterns were further evaluated by stratifying IPD data by comorbidity status.

RESULTS

The relative frequency of serotypes causing IPD differed between adults and children, and also differed between older and younger adults and between adults with and without comorbidities. Serotypes overrepresented as causes of IPD in adults were more commonly carried in older children compared with younger children. In line with this, the serotype-specific frequency of carriage in older children, rather than infants, best correlated with serotype-specific IPD in adults.

CONCLUSIONS

These analyses demonstrate that the serotype patterns in carriage in older children, rather than infants, are best correlated with disease patterns in adults. This might suggest these older children are more influential for disease patterns in adults. These insights could help in optimizing vaccination strategies to reduce disease burden across all ages.

Mobile Gene Sequence Evolution within Individual Human Gut Microbiomes Is Better Explained by Gene-Specific Than Host-Specific Selective Pressures

Pangenomes—the cumulative set of genes encoded by a population or species—arise from the interplay of horizontal gene transfer, drift, and selection. The balance of these forces in shaping pangenomes has been debated, and studies to date focused on ancient evolutionary time scales have suggested that pangenomes generally confer niche adaptation to their bacterial hosts. To shed light on pangenome evolution on shorter evolutionary time scales, we inferred the selective pressures acting on mobile genes within individual human microbiomes from 176 Fiji islanders. We mapped metagenomic sequence reads to a set of known mobile genes to identify single nucleotide variants (SNVs) and calculated population genetic metrics to infer deviations from a neutral evolutionary model. We found that mobile gene sequence evolution varied more by gene family than by human social attributes, such as household or village. Patterns of mobile gene sequence evolution could be qualitatively recapitulated with a simple evolutionary simulation without the need to invoke the adaptive value of mobile genes to either bacterial or human hosts. These results stand in contrast with the apparent adaptive value of pangenomes over longer evolutionary time scales. In general, the most highly mobile genes (i.e., those present in more distinct bacterial host genomes) tend to have higher metagenomic read coverage and an excess of low-frequency SNVs, consistent with their rapid spread across multiple bacterial species in the gut. However, a subset of mobile genes—including those involved in defense mechanisms and secondary metabolism—showed a contrasting signature of intermediate-frequency SNVs, indicating species-specific selective pressures or negative frequency-dependent selection on these genes. Together, our evolutionary models and population genetic data show that gene-specific selective pressures predominate over human or bacterial host-specific pressures during the relatively short time scales of a human lifetime.

Frequency-Dependent Competition Between Strains Imparts Persistence to Perturbations in a Model of Plasmodium falciparum Malaria Transmission

In high-transmission endemic regions, local populations of Plasmodium falciparum exhibit vast diversity of the var genes encoding its major surface antigen, with each parasite comprising multiple copies from this diverse gene pool. This strategy to evade the immune system through large combinatorial antigenic diversity is common to other hyperdiverse pathogens. It underlies a series of fundamental epidemiological characteristics, including large reservoirs of transmission from high prevalence of asymptomatics and long-lasting infections. Previous theory has shown that negative frequency-dependent selection (NFDS) mediated by the acquisition of specific immunity by hosts structures the diversity of var gene repertoires, or strains, in a pattern of limiting similarity that is both non-random and non-neutral. A combination of stochastic agent-based models and network analyses has enabled the development and testing of theory in these complex adaptive systems, where assembly of local parasite diversity occurs under frequency-dependent selection and large pools of variation. We show here the application of these approaches to theory comparing the response of the malaria transmission system to intervention when strain diversity is assembled under (competition-based) selection vs. a form of neutrality, where immunity depends only on the number but not the genetic identity of previous infections. The transmission system is considerably more persistent under NFDS, exhibiting a lower extinction probability despite comparable prevalence during intervention. We explain this pattern on the basis of the structure of strain diversity, in particular the more pronounced fraction of highly dissimilar parasites. For simulations that survive intervention, prevalence under specific immunity is lower than under neutrality, because the recovery of diversity is considerably slower than that of prevalence and decreased var gene diversity reduces parasite transmission. A Principal Component Analysis of network features describing parasite similarity reveals that despite lower overall diversity, NFDS is quickly restored after intervention constraining strain structure and maintaining patterns of limiting similarity important to parasite persistence. Given the described enhanced persistence under perturbation, intervention efforts will likely require longer times than the usual practice to eliminate P. falciparum populations. We discuss implications of our findings and potential analogies for ecological communities with non-neutral assembly processes involving frequency-dependence.

Global genomic pathogen surveillance to inform vaccine strategies: a decade-long expedition in pneumococcal genomics

Vaccines are powerful agents in infectious disease prevention but often designed to protect against some strains that are most likely to spread and cause diseases. Most vaccines do not succeed in eradicating the pathogen and thus allow the potential emergence of vaccine evading strains. As with most evolutionary processes, being able to capture all variations across the entire genome gives us the best chance of monitoring and understanding the processes of vaccine evasion. Genomics is being widely adopted as the optimum approach for pathogen surveillance with the potential for early and precise identification of high-risk strains. Given sufficient longitudinal data, genomics also has the potential to forecast the emergence of such strains enabling immediate or pre-emptive intervention. In this review, we consider the strengths and challenges for pathogen genomic surveillance using the experience of the Global Pneumococcal Sequencing (GPS) project as an early example. We highlight the multifaceted nature of genome data and recent advances in genome-based tools to extract useful information relevant to inform vaccine strategies and treatment options. We conclude with future perspectives for genomic pathogen surveillance.

Negative frequency-dependent selection and asymmetrical transformation stabilise multi-strain bacterial population structures

Streptococcus pneumoniae can be divided into many strains, each a distinct set of isolates sharing similar core and accessory genomes, which co-circulate within the same hosts. Previous analyses suggested the short-term vaccine-associated dynamics of S. pneumoniae strains may be mediated through multi-locus negative frequency-dependent selection (NFDS), which maintains accessory loci at equilibrium frequencies. Long-term simulations demonstrated NFDS stabilised clonally-evolving multi-strain populations through preventing the loss of variation through drift, based on polymorphism frequencies, pairwise genetic distances and phylogenies. However, allowing symmetrical recombination between isolates evolving under multi-locus NFDS generated unstructured populations of diverse genotypes. Replication of the observed data improved when multi-locus NFDS was combined with recombination that was instead asymmetrical, favouring deletion of accessory loci over insertion. This combination separated populations into strains through outbreeding depression, resulting from recombinants with reduced accessory genomes having lower fitness than their parental genotypes. Although simplistic modelling of recombination likely limited these simulations' ability to maintain some properties of genomic data as accurately as those lacking recombination, the combination of asymmetrical recombination and multi-locus NFDS could restore multi-strain population structures from randomised initial populations. As many bacteria inhibit insertions into their chromosomes, this combination may commonly underlie the co-existence of strains within a niche.

Mechanisms That Shape Microbial Pangenomes

Analyses of multiple whole-genome sequences from the same species have revealed that differences in gene content can be substantial, particularly in prokaryotes. Such variation has led to the recognition of pangenomes, the complete set of genes present in a species - consisting of core genes, present in all individuals, and accessory genes whose presence is variable. Questions now arise about how pangenomes originate and evolve. We describe how gene content variation can arise as a result of the combination of several processes, including random drift, selection, gain/loss balance, and the influence of ecological and epistatic interactions. We believe that identifying the contributions of these processes to pangenomes will need novel theoretical approaches and empirical data.

Global genomic epidemiology of Streptococcus pyogenes

Streptococcus pyogenes is one of the Top 10 human infectious disease killers worldwide causing a range of clinical manifestations in humans. Colonizing a range of ecological niches within its sole host, the human, is key to the ability of this opportunistic pathogen to cause direct and post-infectious manifestations. The expansion of genome sequencing capabilities and data availability over the last decade has led to an improved understanding of the evolutionary dynamics of this pathogen within a global framework where epidemiological relationships and evolutionary mechanisms may not be universal. This review uses the recent publication by Davies et al., 2019 as an updated global framework to address S. pyogenes population genomics, highlighting how genomics is being used to gain new insights into evolutionary processes, transmission pathways, and vaccine design.

Frequency-dependent selection can forecast evolution in Streptococcus pneumoniae

Predicting how pathogen populations will change over time is challenging. Such has been the case with Streptococcus pneumoniae, an important human pathogen, and the pneumococcal conjugate vaccines (PCVs), which target only a fraction of the strains in the population. Here, we use the frequencies of accessory genes to predict changes in the pneumococcal population after vaccination, hypothesizing that these frequencies reflect negative frequency-dependent selection (NFDS) on the gene products. We find that the standardized predicted fitness of a strain, estimated by an NFDS-based model at the time the vaccine is introduced, enables us to predict whether the strain increases or decreases in prevalence following vaccination. Further, we are able to forecast the equilibrium post-vaccine population composition and assess the invasion capacity of emerging lineages. Overall, we provide a method for predicting the impact of an intervention on pneumococcal populations with potential application to other bacterial pathogens in which NFDS is a driving force.

Immune exclusion by naturally acquired secretory IgA against pneumococcal pilus-1

Successful infection by mucosal pathogens requires overcoming the mucus barrier. To better understand this key step, we performed a survey of the interactions between human respiratory mucus and the human pathogen Streptococcus pneumoniae. Pneumococcal adherence to adult human nasal fluid was seen only by isolates expressing pilus-1. Robust binding was independent of pilus-1 adhesive properties but required Fab-dependent recognition of RrgB, the pilus shaft protein, by naturally acquired secretory IgA (sIgA). Pilus-1 binding by specific sIgA led to bacterial agglutination, but adherence required interaction of agglutinated pneumococci and entrapment in mucus particles. To test the effect of these interactions in vivo, pneumococci were preincubated with human sIgA before intranasal challenge in a mouse model of colonization. sIgA treatment resulted in rapid immune exclusion of pilus-expressing pneumococci. Our findings predict that immune exclusion would select for nonpiliated isolates in individuals who acquired RrgB-specific sIgA from prior episodes of colonization with piliated strains. Accordingly, genomic data comparing isolates carried by mothers and their children showed that mothers are less likely to be colonized with pilus-expressing strains. Our study provides a specific example of immune exclusion involving naturally acquired antibody in the human host, a major factor driving pneumococcal adaptation.

Predicting the future distribution of antibiotic resistance using time series forecasting and geospatial modelling

Background: Increasing antibiotic resistance in a location may be mitigated by changes in treatment policy, or interventions to limit transmission of resistant bacteria. Therefore, accurate forecasting of the distribution of antibiotic resistance could be advantageous. Two previously published studies addressed this, but neither study compared alternative forecasting algorithms or considered spatial patterns of resistance spread. Methods: We analysed data describing the annual prevalence of antibiotic resistance per country in Europe from 2012 – 2016, and the quarterly prevalence of antibiotic resistance per clinical commissioning group in England from 2015 – 2018. We combined these with data on rates of possible covariates of resistance. These data were used to compare the previously published forecasting models, with other commonly used forecasting models, including one geospatial model. Covariates were incorporated into the geospatial model to assess their relationship with antibiotic resistance. Results: For the European data, which was recorded on a coarse spatiotemporal scale, a naïve forecasting model was consistently the most accurate of any of the forecasting models tested. The geospatial model did not improve on this accuracy. However, it did provide some evidence that antibiotic consumption can partially explain the distribution of resistance. The English data were aggregated at a finer scale, and expected-trend-seasonal (ETS) forecasts were the most accurate. The geospatial model did not significantly improve upon the median accuracy of the ETS model, but it appeared to be less sensitive to noise in the data, and provided evidence that rates of antibiotic prescription and bacteraemia are correlated with resistance. Conclusion: Annual, national-level surveillance data appears to be insufficient for fitting accurate antibiotic resistance forecasting models, but there is evidence that data collected at a finer spatiotemporal scale could be used to improve forecast accuracy. Additionally, incorporating antibiotic prescription or consumption data into the model could improve the predictive accuracy.

Producing polished prokaryotic pangenomes with the Panaroo pipeline

Population-level comparisons of prokaryotic genomes must take into account the substantial differences in gene content resulting from horizontal gene transfer, gene duplication and gene loss. However, the automated annotation of prokaryotic genomes is imperfect, and errors due to fragmented assemblies, contamination, diverse gene families and mis-assemblies accumulate over the population, leading to profound consequences when analysing the set of all genes found in a species. Here, we introduce Panaroo, a graph-based pangenome clustering tool that is able to account for many of the sources of error introduced during the annotation of prokaryotic genome assemblies. Panaroo is available at https://github.com/gtonkinhill/panaroo .

Improved Prediction of Bacterial Genotype-Phenotype Associations Using Interpretable Pangenome-Spanning Regressions

Discovery of genetic variants underlying bacterial phenotypes and the prediction of phenotypes such as antibiotic resistance are fundamental tasks in bacterial genomics. Genome-wide association study (GWAS) methods have been applied to study these relations, but the plastic nature of bacterial genomes and the clonal structure of bacterial populations creates challenges. We introduce an alignment-free method which finds sets of loci associated with bacterial phenotypes, quantifies the total effect of genetics on the phenotype, and allows accurate phenotype prediction, all within a single computationally scalable joint modeling framework. Genetic variants covering the entire pangenome are compactly represented by extended DNA sequence words known as unitigs, and model fitting is achieved using elastic net penalization, an extension of standard multiple regression. Using an extensive set of state-of-the-art bacterial population genomic data sets, we demonstrate that our approach performs accurate phenotype prediction, comparable to popular machine learning methods, while retaining both interpretability and computational efficiency. Compared to those of previous approaches, which test each genotype-phenotype association separately for each variant and apply a significance threshold, the variants selected by our joint modeling approach overlap substantially.IMPORTANCE Being able to identify the genetic variants responsible for specific bacterial phenotypes has been the goal of bacterial genetics since its inception and is fundamental to our current level of understanding of bacteria. This identification has been based primarily on painstaking experimentation, but the availability of large data sets of whole genomes with associated phenotype metadata promises to revolutionize this approach, not least for important clinical phenotypes that are not amenable to laboratory analysis. These models of phenotype-genotype association can in the future be used for rapid prediction of clinically important phenotypes such as antibiotic resistance and virulence by rapid-turnaround or point-of-care tests. However, despite much effort being put into adapting genome-wide association study (GWAS) approaches to cope with bacterium-specific problems, such as strong population structure and horizontal gene exchange, current approaches are not yet optimal. We describe a method that advances methodology for both association and generation of portable prediction models.

Epidemiological consequences of enduring strain-specific immunity requiring repeated episodes of infection

Group A Streptococcus (GAS) skin infections are caused by a diverse array of strain types and are highly prevalent in disadvantaged populations. The role of strain-specific immunity in preventing GAS infections is poorly understood, representing a critical knowledge gap in vaccine development. A recent GAS murine challenge study showed evidence that sterilising strain-specific and enduring immunity required two skin infections by the same GAS strain within three weeks. This mechanism of developing enduring immunity may be a significant impediment to the accumulation of immunity in populations. We used an agent-based mathematical model of GAS transmission to investigate the epidemiological consequences of enduring strain-specific immunity developing only after two infections with the same strain within a specified interval. Accounting for uncertainty when correlating murine timeframes to humans, we varied this maximum inter-infection interval from 3 to 420 weeks to assess its impact on prevalence and strain diversity, and considered additional scenarios where no maximum inter-infection interval was specified. Model outputs were compared with longitudinal GAS surveillance observations from northern Australia, a region with endemic infection. We also assessed the likely impact of a targeted strain-specific multivalent vaccine in this context. Our model produced patterns of transmission consistent with observations when the maximum inter-infection interval for developing enduring immunity was 19 weeks. Our vaccine analysis suggests that the leading multivalent GAS vaccine may have limited impact on the prevalence of GAS in populations in northern Australia if strain-specific immunity requires repeated episodes of infection. Our results suggest that observed GAS epidemiology from disease endemic settings is consistent with enduring strain-specific immunity being dependent on repeated infections with the same strain, and provide additional motivation for relevant human studies to confirm the human immune response to GAS skin infection.

Group A Streptococcus (GAS) is a ubiquitous bacterial pathogen that exists in many distinct strains, and is a major cause of death and disability globally. Vaccines against GAS are under development, but their effective use will require better understanding of how immunity develops following infection. Evidence from an animal model of skin infection suggests that the generation of enduring strain-specific immunity requires two infections by the same strain within a short time frame. It is not clear if this mechanism of immune development operates in humans, nor how it would contribute to the persistence of GAS in populations and affect vaccine impact. We used a mathematical model of GAS transmission, calibrated to data collected in an Indigenous Australian community, to assess whether this mechanism of immune development is consistent with epidemiological observations, and to explore its implications for the impact of a vaccine. We found that it is plausible that repeat infections are required for the development of immunity in humans, and illustrate the difficulties associated with achieving sustained reductions in disease prevalence with a vaccine.

Horizontal gene transfer rate is not the primary determinant of observed antibiotic resistance frequencies in Streptococcus pneumoniae

The extent to which evolution is constrained by the rate at which horizontal gene transfer (HGT) allows DNA to move between genetic lineages is an open question, which we address in the context of antibiotic resistance in Streptococcus pneumoniae. We analyze microbiological, genomic, and epidemiological data from the largest-to-date sequenced pneumococcal carriage study in 955 infants from a refugee camp on the Thailand-Myanmar border. Using a unified framework, we simultaneously test prior hypotheses on rates of HGT and a key evolutionary covariate (duration of carriage) as determinants of resistance frequencies. We conclude that in this setting, there is little evidence of HGT playing a major role in determining resistance frequencies. Instead, observed resistance frequencies are best explained as the outcome of selection acting on a pool of variants, irrespective of the rate at which resistance determinants move between genetic lineages.

Visualizing variation within Global Pneumococcal Sequence Clusters (GPSCs) and country population snapshots to contextualize pneumococcal isolates

Knowledge of pneumococcal lineages, their geographic distribution and antibiotic resistance patterns, can give insights into global pneumococcal disease. We provide interactive bioinformatic outputs to explore such topics, aiming to increase dissemination of genomic insights to the wider community, without the need for specialist training. We prepared 12 country-specific phylogenetic snapshots, and international phylogenetic snapshots of 73 common Global Pneumococcal Sequence Clusters (GPSCs) previously defined using PopPUNK, and present them in Microreact. Gene presence and absence defined using Roary, and recombination profiles derived from Gubbins are presented in Phandango for each GPSC. Temporal phylogenetic signal was assessed for each GPSC using BactDating. We provide examples of how such resources can be used. In our example use of a country-specific phylogenetic snapshot we determined that serotype 14 was observed in nine unrelated genetic backgrounds in South Africa. The international phylogenetic snapshot of GPSC9, in which most serotype 14 isolates from South Africa were observed, highlights that there were three independent sub-clusters represented by South African serotype 14 isolates. We estimated from the GPSC9-dated tree that the sub-clusters were each established in South Africa during the 1980s. We show how recombination plots allowed the identification of a 20 kb recombination spanning the capsular polysaccharide locus within GPSC97. This was consistent with a switch from serotype 6A to 19A estimated to have occured in the 1990s from the GPSC97-dated tree. Plots of gene presence/absence of resistance genes (tet, erm, cat) across the GPSC23 phylogeny were consistent with acquisition of a composite transposon. We estimated from the GPSC23-dated tree that the acquisition occurred between 1953 and 1975. Finally, we demonstrate the assignment of GPSC31 to 17 externally generated pneumococcal serotype 1 assemblies from Utah via Pathogenwatch. Most of the Utah isolates clustered within GPSC31 in a USA-specific clade with the most recent common ancestor estimated between 1958 and 1981. The resources we have provided can be used to explore to data, test hypothesis and generate new hypotheses. The accessible assignment of GPSCs allows others to contextualize their own collections beyond the data presented here.

The research and development process for multiscale models of infectious disease systems

Multiscale modelling of infectious disease systems falls within the domain of complexity science—the study of complex systems. However, what should be made clear is that current progress in multiscale modelling of infectious disease dynamics is still as yet insufficient to present it as a mature sub-discipline of complexity science. In this article we present a methodology for development of multiscale models of infectious disease systems. This methodology is a set of partially ordered research and development activities that result in multiscale models of infectious disease systems built from different scientific approaches. Therefore, the conclusive result of this article is a methodology to design multiscale models of infectious diseases. Although this research and development process for multiscale models cannot be claimed to be unique and final, it constitutes a good starting point, which may be found useful as a basis for further refinement in the discourse for multiscale modelling of infectious disease dynamics.

Complex systems such as infectious disease systems are inherently multilevel and multiscale systems. The study of such complex systems is called complexity science. In this article we present a methodology to design multiscale models of infectious disease systems from a complex systems perspective. Within this perspective we define complexity science as the study of the interconnected relationships of the levels and scales of organization of a complex system. We therefore, define the degree of complexity of a complex system as the number of levels and scales of organization of the complex system needed to describe it. In this work we first present a common multiscale vision of the multilevel and multiscale structure of infectious disease systems as complex systems in which the levels and scales of organization of an infectious disease system interact through different self-sustained multiscale cycles/loops (primary multiscale loops, or secondary multiscale loops, or tertiary multiscale loops) formed at different levels of organization of an infectious disease system due to ongoing reciprocal influence between the microscale and the macroscale. Guided by this multiscale vision, we propose a four-stage research and development process that result in multiscale models of infectious disease systems built from different scientific approaches.