Abstract Predicting the outcome of species or pathogen strain competition is a fundamental aim in both community ecology and infectious disease dynamics. Recent work revealed major challenges in predicting strain co-circulation from ecological coexistence theory due to overcompensatory competition among pathogens for susceptible resources, which can prevent the re-invasion of other competing strains. This resource overcompensation is ubiquitous across host-pathogen systems, but not apparent in simple Lotka-Volterra competition system, highlighting fundamental differences between pathogen strain and species competition. To address this gap, we begin by deriving classical models of pathogen strain and species competition from a resource-consumer model. This generalization illustrates that the relative time scale between resource and consumer dynamics limits the degree of resource overcompensation and therefore dictates the outcome of stochastic competition. Moreover, by introducing a mathematical framework for quantifying pairwise and higher-order terms from general competition systems, we show that a simple, ecological competition model can accurately predict the equilibrium dynamics of strain competition. A case study of rotavirus strain competition reveals that the ability to predict the outcome of strain competition from ecological theory depends on the underlying cross immunity structure. This work synthesizes coexistence theory across two fields by providing a unifying framework for predicting the outcome of complex ecological competition.

Foundation models-large AI systems pretrained on broad, heterogeneous data-are transforming scientific discovery. These models (e.g., GPT, GenCast, AlphaFold) excel at learning generalizable representations and adapting to new tasks with limited data. Yet, epidemic modeling has not experienced a comparable transformation. Traditional models remain pathogen-specific and often struggle to generate rapid insights during emerging outbreaks, as starkly illustrated by the SARS-CoV-2 pandemic. This Perspective asks whether the foundation model paradigm can extend to epidemic science: Can we build a single, pretrained model that captures the shared principles of infectious disease dynamics across pathogens, populations, and settings? Such a model could be fine-tuned to new contexts with minimal data, enabling faster forecasting, inference, and response, especially valuable in resource-limited settings. We argue that the growing convergence of epidemiological insight and modern AI makes this goal both urgent and increasingly plausible. We outline the main challenges in building foundation models for epidemics-nonstationarity, fragmented surveillance data, presence of diverse dynamical regimes, and the need for interpretability. We then propose a roadmap toward epidemic foundation models, emphasizing both algorithmic innovations to address these challenges and progress beyond algorithms, including investments in open datasets and cross-disciplinary training and collaboration. Developing epidemic foundation models offers a potentially transformative opportunity to strengthen global health security, particularly by improving preparedness in underresourced settings. If successful, they will serve as powerful, generalizable tools that complement existing efforts. The process of building these models will itself be valuable, exposing critical data gaps and guiding investments in global surveillance.

The increasing interconnectedness of the modern world calls for globally equitable solutions to combat pandemic challenges. However, we have seen a tendency in recent decades for high-income countries to resort to “vaccine nationalism,” hoarding vaccine production to the detriment of lower-income countries. In addition, vaccine nationalism can prove detrimental to hoarding countries in the long term, as inequitable global vaccine distribution during the COVID-19 risked exacerbating the rise of harmful immune-escape variants that largely counteracted the original benefits of vaccine hoarding. Thus, vaccine hoarding may create a problem of time preference for a vaccine-producing country, where countries heavily discounting the future would opt for vaccine hoarding while countries lightly discounting the future would opt for vaccine sharing. Using a novel modeling framework integrating epidemiological, evolutionary, and economic processes, we demonstrate how high temporal discounting, low levels of outgroup prosociality, and high vaccine-distribution costs for low-income countries can promote vaccine-hoarding tendencies. We further show how these factors interact with epidemiological and evolutionary parameters to incentivize vaccine sharing in different ways: in some parameter regimes, vaccine sharing helps by reducing variant infections, while in others, vaccine sharing helps by reducing the probability of initial variant emergence. As a result, the optimal fraction of vaccines a country should share in our model is a bimodal function of the pathogen’s transmissibility. We thus provide a nuanced, model-based exploration of how various factors may contribute to vaccine nationalism’s emergence, emphasizing the need for international organizations to coordinate global vaccination responses to future pandemics.

Abstract The ongoing epizootic of highly pathogenic avian influenza (HPAI) continues to cause massive deaths in wildlife. Fundamental understanding of its disease ecology in natural populations is urgently needed. This knowledge has been hindered by the difficulty of acquiring data on epidemic dynamics. Here, using data collected from a threatened population of Dalmatian pelicans ( Pelecanus crispus ), we recover the epidemiological and evolutionary history of one of the largest HPAI wildlife mortality events. The results show that this devastating outbreak was likely seeded by a single introduction associated with movement of the species. By estimating epidemiological features of two consecutive outbreaks in the same population, we show that panzootic H5N1 since 2022 likely exhibits higher transmissibility and longer shedding time in non-reservoir birds, compared to previous H5NX subtypes. We also evaluate effectiveness of past and future control measures: carcass removal during the outbreak is shown to have surprisingly little impact on mitigating the mortality; and current H5 vaccines relying on capture and injection to deliver cannot establish herd immunity in a wildlife population. The results provide the first field evidence supporting the hypothesis that viral fitness difference of H5N1 to previous H5NX subtypes is the key cause of the expanded epizootic and panzootic since 2022, and on highly debated HPAI management strategies in wildlife populations. Author Summary Since late 2021, a panzootic of H5N1 highly pathogenic avian influenza (HPAI) has caused unprecedented mass mortality in wildlife. Many severely affected species are critical for ecosystem functions, including several threatened and endangered species. However, fundamental knowledge of HPAI disease ecology in natural populations is still lacking, and the effectiveness of potential controls is under debate. Here, using data collected from one of the largest HPAI outbreak in wild animals – over 1700 deaths (80% of the population) in a threatened population of Dalmatian pelicans in Greece, for the first time we recover the transmission dynamics of H5N1 in a migratory bird population. Based on the recovered dynamics, we show that removing carcasses during the outbreak was surprisingly ineffective, and future potential vaccination would require a novel delivery method to establish population immunity in wildlife. Our study provides new insight in the epidemiology of HPAI clade 2.3.4.4b in wildlife, and provides a foundation for assessing interventions within this complicated system.

Defective interfering particles (DIPs) are incomplete viral genomes that modulate infection by competing with wild-type viruses and activating the innate immune response. Activation of the immune response leads to the production of cytokines and chemokines, including type I interferon (IFN), which restricts viral growth and may cause cell death. How DIPs interact with type I interferon (IFN) in spatially structured environments remains unclear. Focusing here on influenza A viruses, we developed a spatially explicit, stochastic model of in vitro viral infection that integrates virus and DIP replication, IFN signalling, and alternative dispersal modes. We find that: (1) our model captures the ring-like and patchy plaque morphologies observed experimentally; (2) IFN production peaks at an intermediate DIP ratio, reflecting a trade-off between early immune activation and sufficient co-infection; and (3) even a small fraction of long-range spread by virus and DIPs enables escape from the immune-based containment despite long-range IFN diffusion; this causes stronger antiviral responses but earlier peaks in virus egress at similar levels of cell loss. The model is available as an interactive platform: https://shiny-spatial-infection-app-production.up.railway.app/.

Abstract Virus exposure history, particularly first exposure, is believed to shape vaccine efficacy and infection susceptibility; however, evidence for mechanistic links between immune responses in individuals and epidemiological outcome in populations is scarce. Recent co-circulation of SARS-CoV-2 variants XFG and BA.3.2 has revealed a striking enrichment in BA.3.2 cases among children. By combining epidemiological modeling, serology and monoclonal antibody analysis in children and adults, we show the dependence of effective variant-specific antibodies on vaccination history which may explain birth-year influence on differential susceptibility to these co-circulating variants. Ancestral cross-reactive site I antibodies frequently neutralize BA.3.2, but not XFG. By contrast, Omicron type-specific site I/III and III antibodies frequently neutralize XFG but not BA.3.2, revealing a tradeoff in the ability to neutralize these two co-circulating strains. These findings mechanistically link immune history, variant neutralization, antibody repertoire and variant infection risk, and suggest that vaccination regimens in children should prioritize neutralization breadth.

BACKGROUND:The COVID-19 pandemic has illustrated how nonpharmaceutical interventions (NPIs), such as mask-wearing, social distancing, and purifying air, can successfully mitigate transmission and reduce infections in the short term. However, the longer-term implications of these interventions on infection levels are less clear. In tandem, recent observational evidence suggests that the relative susceptibility of partially immune individuals to infection may be dose-dependent, i.e., higher exposures are more likely to result in (re)infection in individuals with prior immunity from infection or vaccination. METHODS:To examine this question, we use mechanistic immuno-epidemiological models to determine the equilibrium infection levels with NPI-induced reductions in transmission in the presence of this dose-dependency. RESULTS:We find that NPIs can successfully decrease the number of infections at endemicity, even in high transmission scenarios. We also show that this effect is heightened by vaccination, especially if a durable, broadly-protective (i.e., broad antigen specificity) vaccine is deployed. Finally, we find that NPI-induced declines in infection levels are strongly magnified if the characteristics of subsequent infections, such as transmissibility or duration, are also dose-dependent. CONCLUSIONS:Overall, our results suggest that the long-term usage of NPIs could successfully reduce infections especially where immunity-exposure trade-offs apply due to dose-dependency, and illustrate the urgent need to characterize the underlying relationships between exposure and host immune responses.

Understanding the geographic spread of emerging respiratory viruses is critical for pandemic preparedness, yet the early spatiotemporal dynamics of the 2009 H1N1 pandemic influenza and severe acute respiratory syndrome coronavirus 2 in the United States remain unclear. While mobility and genomic data have revealed important aspects of pandemic spatial spread, several key questions remain: Did the two pandemics follow similar spatial transmission routes? How rapidly did they spread across the United States? What role did stochastic processes play in early spatial transmission? To address these questions, we integrated high-resolution disease data with a robust, data-efficient inference framework combining air travel, commuting flows, and pathogen superspreading potentials to reconstruct their spatial spread across US metropolitan areas. The two pandemics exhibited distinct transmission pathways across locations; however, both pandemics established local circulation in most metropolitan areas within weeks, driven by several shared transmission hubs. Early spatial spread was more strongly associated with air travel than with commuting, though stochastic dynamics introduced substantial uncertainty in transmission routes, creating challenges for timely detection and control. Simulations indicate that broad wastewater surveillance coverage beyond top transmission hubs coupled with effective infection control may slow initial spatial expansion. Our findings highlight the rapid, stochastic spread of pandemic respiratory pathogens and the difficulties of early outbreak containment.

In 1798, Jenner’s smallpox vaccine made it possible to prevent the deadliest of childhood diseases. In Denmark the vaccine was used from 1801, and by 1810 a mandatory 1-dose childhood vaccination program was instituted, free of charge. As proof of vaccination (or natural immunity) was required for church confirmation, most children were vaccinated and smallpox disappeared from Copenhagen after 1810 [[1][1]]. After a 14-year “honeymoon period”, smallpox returned in 1824 with a new face: a milder disease in mostly young adults. Here we investigate the effects of smallpox vaccination on the epidemic patterns throughout the post-honeymoon era (1824-1875). We accessed data from the hospital “Søkvæsthuset” where all smallpox cases, mild and severe, were hospitalized during 1824-1835 in order to contain the outbreak. We identified ∼ 3000 smallpox cases in four separate epidemics occurring during this period and accessed data on the age and vaccination status of patients where available. In addition, some information on the number and age distribution of severe smallpox cases and on mortality for outbreaks in the 1860’s and 70’s is available. We used a mechanistic model (SEIR) to assess factors explaining the return of smallpox, and the changing mean age of cases. We considered vaccination coverage and effectiveness, duration of vaccination-induced immunity, and the fate of the “lost generation” of persons born around 1800, too early to get vaccinated and too late to have been infected with smallpox. Our model tracks well the disappearance and return of smallpox in 1824, its age patterns, and the interval between epidemic peaks. Our study shows that vaccine waning after 25 years on average was likely the primary reason explaining the return of smallpox and the change in its epidemiology in the vaccine era. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement This project received funding from the Danish National Research Foundation (DNRF170) and NordForsk (project no. 104910). ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The Danish National Archives (Rigsarkivet) I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Danish National Research Foundation, https://ror.org/00znyv691, DNRF170 NordForsk, https://ror.org/05bqzfg94, 104910 [1]: #ref-1

Interventions to slow the spread of SARS-CoV-2 significantly disrupted the transmission of other pathogens. As interventions lifted, whether and when human pathogens would eventually return to their prepandemic dynamics remains to be answered. Here, we present a framework for estimating pathogen resilience based on how fast epidemic patterns return to their prepandemic dynamics. By analyzing time series data from Hong Kong, Canada, Korea, and the United States, we quantify the resilience of common respiratory pathogens and further predict when each pathogen will eventually return to its prepandemic dynamics. Our predictions are able to distinguish which pathogens should have returned already, and deviations from these predictions reveal long-term impacts of pandemic perturbations. We find a faster rate of susceptible replenishment underlies pathogen resilience and sensitivity to both large and small perturbations. Overall, our analysis highlights the persistent nature of common respiratory pathogens compared to vaccine-preventable infections, such as measles.

BACKGROUND:The Pandemic Response Repository through Microbial and Immune Surveillance and Epidemiology (PREMISE) programme was established to translate knowledge gained from global immunoepidemiological surveillance into a better understanding of population-level dynamics of emerging and re-emerging infections, as well as into the discovery and development of biomedical countermeasures against potential pandemic threats. As proof of principle for this approach, we conducted a longitudinal immunoepidemiological study in children in the USA, focusing on enterovirus D68 (EV-D68) infection dynamics but also capturing surveillance of a broad array of other endemic respiratory pathogens. Serendipitously, our sampling spanned the lifting of widespread COVID-19 non-pharmaceutical interventions (NPIs) in 2022-23, following a unique period during which virus exposure markedly diminished. METHODS:This prospective, multicentre, longitudinal, immunoepidemiological surveillance study enrolled children aged 10 years or younger and weighing at least 8 kg at three US university sites. Blood specimens collected from January to June, 2022 (visit 1; pre-enterovirus season), and from January to June, 2023 (visit 3; post-enterovirus season), were tested in a multiplex assay for antibody binding to EV-D68 (prespecified primary objective) and a panel of 15 other respiratory viruses (exploratory objectives), and for neutralising activity against EV-D68, enterovirus A71, and respiratory syncytial virus (RSV; for antibody binding assay validation). Respiratory mid-turbinate swabs collected from children with symptomatic illness who participated in symptom surveys during July-December, 2022 (visit 2; enterovirus season), underwent metagenomic sequencing for pathogen detection. Serological data for EV-D68 were incorporated into epidemiological models based on case data from national surveillance to predict future transmission dynamics. FINDINGS:Of 488 eligible children approached, 174, with a median age of 3·4 years (IQR 1·9-6·4), were enrolled and followed up longitudinally from January, 2022, to June, 2023. Three children withdrew before study completion and 51 were lost to follow-up between visits 1 and 3. 90 paired serological samples and 73 respiratory swabs were tested. Mean antibody binding and neutralisation titres against all viruses tested increased over the study period, most notably in younger children with lower initial titres. The highest exposure rates (seroconversion or antibody boosting) were seen with SARS-CoV-2 (51 [59%] of 87), EV-D68 (36 [41%] of 87), RSV (36 [41%] of 87), and influenza (35 [40%] of 87), whereas the pathogens most frequently detected by respiratory swab sequencing were EV-D68 (clade B3), rhinovirus A, and rhinovirus C (n=7 each). Incorporating EV-D68 serological data into epidemiological models resulted in an 82% reduction in the range of prediction errors and a 33% reduction in median prediction errors for longer-term EV-D68 circulation dynamics compared with national pathogen surveillance data alone. INTERPRETATION:In this study, we captured immunological evidence of endemic virus re-emergence in children following lifting of pandemic NPIs, which revealed high rates of exposure to endemic respiratory pathogens in a large group of seronegative, predominantly younger, children. This study demonstrates the feasibility and utility of immunoepidemiological surveillance to enable more precise and accurate modelling of pathogen circulation dynamics to predict and prepare for future waves of disease. FUNDING:Intramural Research Program of the National Institute of Allergy and Infectious Diseases-Vaccine Research Center, and the National Cancer Institute, National Institutes of Health.

Titrating the relative importance of endogenous and exogenous drivers for dynamical transitions in host-pathogen systems remains an important research frontier towards predicting future outbreaks and making public health decisions. In Japan, respiratory syncytial virus (RSV), a major childhood respiratory pathogen, displayed a sudden, dramatic shift in outbreak seasonality (from winter to fall) in 2016. This shift was not observed in any other countries. We use mathematical models to identify processes that could lead to this outcome. In line with previous analyses, we identify a robust quadratic relationship between mean specific humidity and transmission, with minimum transmission occurring at intermediate humidity. This drives semiannual patterns of seasonal transmission rates that peak in summer and winter. Under this transmission regime, a subtle increase in population-level susceptibility can cause a sudden shift in seasonality, where the degree of shift is primarily determined by the interval between the two peaks of seasonal transmission rate. We hypothesize that an increase in children attending childcare facilities may have contributed to the increase in susceptibility through increased contact rates with susceptible hosts. Our analysis underscores the power of studying infectious disease dynamics to titrate the roles of underlying drivers of dynamical transitions in ecology.

Estimating the durability of immunity from vaccination is complicated by unreported revaccination and unobserved natural infection or reexposure, which could result in overestimation of protection longevity. We tested serial, cross-sectional serum samples from 2005 to 2015 (n = 2530) for immunoglobulin G (IgG) to examine measles seroprevalence, spatiotemporal patterns of titers across regions, and antibody dynamics among children aged 1-9 years who grew up during varying measles circulation in Madagascar under a 1-dose vaccination schedule. We found that measles seroprevalence generally decreased over this period. Furthermore, we conducted 2 nested serological surveys, analyzing 393 samples taken in 2005 (n = 158), a time point preceded by high levels of measles circulation, and 2015 (n = 235), a time point preceded by low levels of measles circulation. Among children alive during periods of limited measles circulation, we found lower measles seroprevalence in all age groups and lower antibody titers in children aged 7-9 years old. Notably, titers among children aged 7-9 dipped near the threshold of protection, highlighting the importance of additional measles vaccine doses. Our findings suggest vulnerabilities might emerge during periods of limited measles circulation for countries with a 1-dose schedule, due to both the buildup of susceptible individuals and waning titers.

The Vendi score (VS), a diversity metric recently conceived in the context of machine learning, with applications in a wide range of fields, has a few distinct advantages over the metrics commonly used in ecology. It is classification-independent, incorporates abundance information, and has a tunable sensitivity to rare/abundant types. Using rich COVID-19 sequence data as a paradigm, we develop methods for applying the VS to time-resolved sequence data. We show how the VS allows for characterization of the overall diversity of circulating viruses and for discernment of emerging variants prior to formal identification. Furthermore, applying the VS to phylogenetic trees provides a convenient overview of within-clade diversity which can aid viral variant detection.

The ability of viruses to adapt and evolve as they spread through a population remains a global concern. Immune responses drive viral evolution, but our understanding of how specific selective pressures from distinct mediators of innate and adaptive immune responses within individual hosts influence long-term pathogen evolutionary trajectories is limited. Here, we argue that there is a critical need for experiments that bridge individual host studies on the one hand and population-level epi-evolutionary studies on the other. Current frameworks that investigate how immune parameters individually affect viral abundance or clearance need to be more frequently coupled with sequencing data across viral genomes. Such integration would enable the identification of potentially distinct immune-mediated selection pressures on viruses, thereby better informing the design of future cross-scale experimental models. Resolving how immune factors influence the emergence of novel viral variants will be essential to better predict and effectively manage viral evolution.

Risk-driven behaviour provides a feedback mechanism through which individuals both shape and are collectively affected by an epidemic. We introduce a general and flexible compartmental model to study the effect of heterogeneity in the population with regard to risk tolerance. The interplay between behaviour and epidemiology leads to a rich set of possible epidemic dynamics. Depending on the behavioural composition of the population, we find that increasing heterogeneity in risk tolerance can either increase or decrease the epidemic size. We find that multiple waves of infection can arise due to the interplay between transmission and behaviour, even without the replenishment of susceptibles. We find that increasing protective mechanisms such as the effectiveness of interventions, the fraction of risk-averse people in the population and the duration of intervention usage reduce the epidemic overshoot. When the protection is pushed past a critical threshold, the epidemic dynamics enter an underdamped regime where the epidemic size exactly equals the herd immunity threshold and overshoot is eliminated. Finally, we can find regimes where epidemic size does not monotonically decrease with a population that becomes increasingly risk-averse.

Vaccination against livestock diseases is an effective method to prevent and control the spread of pathogens and reduce antimicrobial consumption and livestock production losses. Systematic data on global vaccination coverage could unlock opportunities to expand these outcomes. In this study, we estimate annual vaccination coverage and disease incidence for 104 cattle, porcine, and poultry diseases in 203 reporting countries and territories between 2005 and 2025 using data from the World Animal Health Information System, the Food and Agriculture Organization, and published literature. We provide 686,559 data points and further evaluate 11 diseases most widely targeted by vaccination programs in 2025. The vaccination coverage for global populations at risk of these diseases in 2025 is as follows: for cattle, 16.64% (95% CI: 16.63 to 16.66) against foot and mouth disease, 33.80% (33.43 to 34.38) against lumpy skin disease, 7.46% (6.71 to 8.81) against Brucella abortus, 11.57% (10.29 to 13.36) against anthrax, and 7.93% (6.27 to 14.09) against rabies. For pigs, 6.56% (6.56 to 6.57) against classical swine fever, 4.96% (3.28 to 8.76) against anthrax, and 8.08% (5.10 to 17.20) against rabies. For poultry, 17.62% (17.37 to 18.04) against Newcastle disease, 16.71% (16.42 to 19.01) against infectious bronchitis, 9.17% (8.59 to 13.67) against infectious laryngotracheitis, 15.04% (14.63 to 18.63) against infectious bursal disease, and 8.81% (7.94 to 11.97) against Marek's disease. Expanding vaccination efforts in India and Argentina for cattle; China and Russia for pigs; and China, Brazil, and Iran for poultry may yield the greatest reductions in global livestock disease burden.

Our understanding of influenza transmission remains imperfect due to the high prevalence of asymptomatic infections that often go undetected. To address this challenge, we leveraged uniquely resolved data from a household cohort study spanning three consecutive years in rural and urban South Africa. The study incorporated pre-season serum collection and twice-weekly virological testing during the influenza season, regardless of symptom presence. We developed a subtype/lineage-specific influenza household transmission model that accounts for time-resolved viral shedding across the full clinical spectrum of infections, allowing us to disentangle the role of household and community exposures, pre-season immunity, and age on transmission. Our analysis revealed that viral shedding intensity, as measured by the cycle threshold (Ct) values of infected household members, significantly correlated with the risk of transmission for all four influenza subtypes/lineages. After adjusting for viral shedding, pre-season hemagglutination inhibition (HAI) titers greater than 1:40 were associated with a significantly lower risk of infection acquisition for A(H1N1)pdm09, A(H3N2), and B/Victoria, but not for B/Yamagata. Notably, children exhibited higher susceptibility, longer viral shedding durations, and higher peak viral loads compared to adults across all subtypes/lineages, even after adjusting for pre-season HAI titers. While our findings support that HAI titers correlate with protection, the strong residual effects of age on susceptibility and viral shedding may reflect the accumulation of additional immune responses shaped by repeated exposures over time. Our study underscores the need to explore immune mechanisms beyond HAI titers that modulate influenza susceptibility and transmission. ### Competing Interest Statement CC has received grant support from Sanofi Pasteur, US CDC, the Bill & Melinda Gates Foundation, the Taskforce for Global Health, Wellcome Trust and the South African Medical Research Council. AvG has received grant support from Sanofi Pasteur, Pfizer related to pneumococcal vaccine, CDC and the Bill & Melinda Gates Foundation. NW reports grants from Sanofi Pasteur and the Bill & Melinda Gates Foundation. NAM has received a grant to his institution from Pfizer to conduct research in patients with pneumonia and from Roche to collect specimens to assess a novel TB assay. ### Funding Statement Data collection of the PHIRST flu study was supported by the National Institute for Communicable Diseases of the National Health Laboratory Service and the US Centers for Disease Control and Prevention (cooperative agreement number 5U51IP000155). Molly Sauter would like to acknowledge support from Princeton University Office of Undergraduate Research Undergraduate Fund for Academic Conferences through the President′ Fund. Bryan Grenfell would like to acknowledge support from Princeton Catalysis and Princeton Precision Health. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The PHIRST protocol was approved by the University of the Witwatersrand Human Research Ethics Committee (Reference 150808) and the US CDC's Institutional Review Board relied on the local review (#6840). The protocol was registered on http://clinicaltrials.gov on 6 August 2015 (Reference [NCT02519803][1]). Participants provided individual written consent or assent prior to enrollment and received a grocery store voucher of ZAR25–30 (USD 2–2.5) per visit for their time. (See 45 C.F.R. part 46.114; 21 C.F.R. part 56.114). I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes The study protocol including informed consent forms is available on the NICD website (https://www.nicd.ac.za/wp-content/uploads/2021/02/PHIRST-SARS-CoV-2-protocol-V1-amendment-Nov2020-incl-upd-consent.pdf). The investigators welcome enquiries about possible collaborations and requests for access to the data set. Data will be shared after approval of a proposal and with a signed data access agreement. Investigators interested in more details about this study, or in accessing these resources, should contact the principle investigator, Prof Cheryl Cohen, at NICD (cherylc@nicd.ac.za). [1]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT02519803&atom=%2Fmedrxiv%2Fearly%2F2025%2F03%2F26%2F2025.03.25.25324622.atom

The epidemiological dynamics of Mycoplasma pneumoniae is characterized by poorly understood complex multiannual cycles. The origins of these cycles have long been debated, and multiple explanations of varying complexity have been suggested. Using Bayesian methods, we fit a dynamical model to half a century of M. pneumoniae surveillance data from Denmark (1958 to 1995, 2010 to 2025) and uncover a parsimonious explanation for the persistent cycles, based on the theory of quasicycles. The period of the multiannual cycle (approx. 5 y in Denmark) is explained by susceptible replenishment due, primarily, to loss of immunity. While an excellent fit to shorter time series (a few decades), the deterministic model eventually settles into an annual cycle, unable to reproduce the persistent cycles. We find that environmental stochasticity (e.g., varying contact rates) stabilizes the multiannual cycles and so does demographic noise, at least in smaller or incompletely mixing populations. The temporary disappearance of cycles during 1979 to 1985 is explained as a consequence of stochastic mode-hopping. The circulation of M. pneumoniae was recently disrupted by COVID-19 nonpharmaceutical interventions (NPIs), providing a natural experiment on the effects of large perturbations. Consequently, the effects of NPIs are included in the model and medium-term predictions are explored. Our findings highlight the intrinsic sensitivity of M. pneumoniae dynamics to perturbations and interventions, underscoring the limitations for long-term prediction. More generally, our findings provide further evidence for the role of stochasticity as a driver of complex cycles across endemic and recurring pathogens.

While excess rainfall is associated with mosquito-borne disease because it supports mosquito breeding, drought may also counterintuitively increase disease transmission by altering mosquito and host behavior. This phenomenon is important to understand because climate change is projected to increase both extreme rainfall and drought. In this study, we investigated the extent to which seasonally-driven mosquito and primate behavior drove the first yellow fever virus (YFV) epidemic in an urban area in Brazil in nearly a century, coinciding with an equally rare drought, and to assess the role of interventions in ending the outbreak. We hypothesized that drought triggered the outbreak by driving the forest mosquitoes and non-human primates towards the city in search of water and that the mosquitoes were biting more frequently to avoid desiccation. A dynamical YFV model supports these hypotheses, showing that both behavioral changes were needed to explain the outbreak timing and incidence. Further, a combination of vector control, conservation measures, and vaccination contributed to ending the outbreak, with the strongest effects from vaccination. Together, these results suggest that drought, likely to become more frequent in this region in the coming decades, can significantly influence mosquito-borne disease transmission, and that sustained control will require multiple interventions.