Overall, we imputed a total of 10,558 (posterior median; 95% CrI: 10,394C10,750) unique infections (observe Post-processing of illness history posteriors in Materials and Methods) across all individuals and times from your Fluscape data and 547 (posterior median; 95% CrI: 519C574) infections from your Ha Nam, Viet Nam data

Overall, we imputed a total of 10,558 (posterior median; 95% CrI: 10,394C10,750) unique infections (observe Post-processing of illness history posteriors in Materials and Methods) across all individuals and times from your Fluscape data and 547 (posterior median; 95% CrI: 519C574) infections from your Ha Nam, Viet Nam data. assay measurements against 20 A/H3N2 strains from repeat serum samples collected between 2009 and 2015. These estimated illness histories allowed us to reconstruct historic seasonal influenza patterns and to investigate how influenza incidence varies over time, space and age with this populace. We estimated median annual influenza illness rates to be approximately 18% from 1968 to 2015, but with considerable variance between years. 88% of individuals were estimated to have been infected at least once during the study period (2009C2015), and 20% were estimated to have three or more infections in that time. We inferred decreasing contamination rates with increasing age, and found that annual attack rates were highly correlated across all locations, regardless of their distance, suggesting that age has a stronger impact than fine-scale spatial effects in determining an individuals antibody profile. Finally, we reconstructed each individuals expected antibody profile over their lifetime and inferred an age-stratified relationship between probability of contamination and HI titre. Our analyses show how multi-strain serological panels provide rich information on long term, epidemiological trends, within-host processes and immunity when analyzed using appropriate inference methods, and adds to our understanding of the life course epidemiology of influenza A/H3N2. Introduction Patterns of influenza infections in humans are highly varied across time, space and demography [1,2]. Recurrent epidemics occur because influenza viruses undergo an evolutionary process of antigenic drift, whereby new strains escape pre-existing host immunity through the accumulation of mutations in JAK/HDAC-IN-1 immunodominant surface glycoproteins leading to rapid turnover of lineages, with specific strains persisting for 1C2 years [3,4]. Because individuals are alive at different times and locations, they are exposed to different strains and thus each individual has a distinct immunological history [5,6]. As a result, serological data suggest that humans are infected with a new A/H3N2 influenza strain approximately every 5 years, with less frequent infections, or at least less frequent detectable antibody boosts, as individuals enter middle age [7,8]. A better understanding of who, where and when influenza infections are likely to occur would aid in public health planning, nowcasting and forecasting [9,10]. However, it is not just antigenic variation and evolution that contributes to variation in influenza incidence, but a combination of individual and population level factors [11,12]. Birth cohorts [13C15], contact and movement patterns [16C18], climatic variation [19,20], school terms [21,22], city JAK/HDAC-IN-1 structure [23,24], and household structure [25,26] have all been shown to be associated with variation in influenza incidence. However, variation in surveillance quality and consistency across locations and over time makes it difficult to identify individual-level or population-specific effects over a longer time period using routine influenza-like-illness surveillance data [27,28]. These limitations may be overcome by using serological data, where unobserved past infections and vaccinations leave a signature in an individuals measurable antibody profile [29C31]. For influenza, measured antibody levels are the result of complex interactions of immunological responses from all past exposures [6,32]. Hence, accurate inferences of individual contamination histories require models of antibody kinetics to determine the number and timing of past exposures to multiple influenza strains [8,13,33C35]. These models can be complicated, as immunological interactions of antigenic drift with immune memory occur through imprinting effects, whereby the set and order of strains in an individuals previous exposure history influences which epitopes are targeted and the magnitude of their antibody response to subsequent exposures [6,32]. Estimating influenza contamination histories MOBK1B from serological data therefore presents a decoding problem, as the space of possible exposure histories which could lead to an observed antibody landscape is usually large, and observed antibody titres are highly variable due to within-host and JAK/HDAC-IN-1 laboratory-level effects. Although inferences which account for these mechanisms have provided rich insights into individual-level life course immune profiles, JAK/HDAC-IN-1 most attempts have been in relatively small cohorts or using small panels of influenza strains, limiting the conclusions which can be drawn about population-level influenza epidemiology [13,36,37]. Here, we applied an infection history inference method to data from a large serosurvey to reconstruct lifetime individual contamination histories and population-level incidence of A/H3N2 influenza in Guangzhou, China [35,36,38]. Contamination histories were inferred based on individual-level antibody profiles to a panel of 20 influenza A/H3N2 strains representing viruses JAK/HDAC-IN-1 that first circulated from 1968 onward. The study population comes from a range of age groups, social backgrounds, and geographical areas, thereby providing an ideal dataset to investigate predictors of influenza contamination and small-scale spatial variation. In fitting the model, we also obtained parameter estimates for the underlying antibody kinetics model which allows us to elucidate.