Modeling infection transmission in primate networks to predict centrality‐based risk

Abstract

Social structure can theoretically regulate disease risk by mediating exposure to pathogens via social proximity and contact. Investigating the role of central individuals within a network may help predict infectious agent transmission as well as implement disease control strategies, but little is known about such dynamics in real primate networks. We combined social network analysis and a modeling approach to better understand transmission of a theoretical infectious agent in wild Japanese macaques, highly social animals which form extended but highly differentiated social networks. We collected focal data from adult females living on the islands of Koshima and Yakushima, Japan. Individual identities as well as grooming networks were included in a Markov graph-based simulation. In this model, the probability that an individual will transmit an infectious agent depends on the strength of its relationships with other group members. Similarly, its probability of being infected depends on its relationships with already infected group members. We correlated: (i) the percentage of subjects infected during a latency-constrained epidemic; (ii) the mean latency to complete transmission; (iii) the probability that an individual is infected first among all group members; and (iv) each individual's mean rank in the chain of transmission with different individual network centralities (eigenvector, strength, betweenness). Our results support the hypothesis that more central individuals transmit infections in a shorter amount of time and to more subjects but also become infected more quickly than less central individuals. However, we also observed that the spread of infectious agents on the Yakushima network did not always differ from expectations of spread on random networks. Generalizations about the importance of observed social networks in pathogen flow should thus be made with caution, since individual characteristics in some real world networks appear less relevant than they are in others in predicting disease spread. Am. J. Primatol. 78:767-779, 2016. © 2016 Wiley Periodicals, Inc.

The influence of phylogeny, social style, and sociodemographic factors on macaque social network structure

“… b Calculated as the mean observation time per subject in hours, for datasets collected using focal animal (mean focal observation time), all‐occurrences (total observation time of the group), and/or scan sampling (number of scans per subject times the duration of the scan (derived from Romano et al ()). Blank entries represent datasets in which ad‐libitum sampling was used.…”

“…A minority used either scan sampling (aggression: 2/38, or 5%; grooming: 7/34, or 20%) or ad‐libitum sampling along with focal animal sampling (aggression: 3/38, or 8%; grooming: 2/34, or 6%). The scan sampling datasets were included based on a previous study on Japanese macaques revealing that for a given number of scans, this approach yields identical amounts of behavioral data to focal sampling for the same duration (in minutes) as the number of scans (Romano et al, ). Datasets that used ad libitum sampling were included after ensuring that they had been (a) used in previous comparative studies (Balasubramaniam et al, ; Schino & Aureli, ), and (b) conducted for overall durations that were comparable to those of the focal‐ or all‐occurrences sampled datasets (6–12 months).…”

Primate Infectious Disease Ecology: Insights and Future Directions at the Human-Macaque Interface

“…Second, studies on (2) primate parasite socioecology, in addition to the relative role(s) of resource abundance, predation pressure, and infanticidal risk, have also begun to examine the role of parasites in shaping the evolution of primate group sizes and social network structure Nunn et al 2011; meta-analyses by Griffin and Nunn 2012;Nunn et al 2015;Patterson and Ruckstuhl 2013;Rifkin et al 2012). Conversely, the idea that group-living and social structure may also impact the diversity and prevalence of parasites in hosts (Drewe and Perkins 2015;VanderWaal and Ezenwa 2016) has led to such socioecological approaches to also focus on the identification of potential "super spreaders" or "social bottlenecks" of infection (Balasubramaniam et al , 2018Duboscq et al 2016;Griffin and Nunn 2012;MacIntosh et al 2012;Romano et al 2016). Other studies have used agent-based models to predict the prevalence and transmission of parasites through artificial primate groups and networks (Griffin and Nunn 2012;Nunn et al 2015).…”

“…Among all the primates, macaques (particularly rhesus macaques and long-tailed macaques) continue to be the most common genus used in captivity as models for biomedical research (Hannibal et al 2017;Phillips et al 2014). Further, as we review above, infectious disease research among free-living macaque populations have had variant foci, ranging from the detection and diagnosis of parasites (Engel and Jones-Engel 2011;Engel et al 2008;Jones-Engel et al 2004, through establishing co-evolutionary links between parasites and macaque hosts (Huffman et al 2013b), to assessing the social and environmental underpinnings of parasite prevalence and transmission MacIntosh et al 2010MacIntosh et al , 2012Romano et al 2016). The OH concept, along with our above-stated argument that human-macaque interfaces are functionally interdependent, coupled systems, would provide a means to bring such diverse foci under a single, unifying framework (Destoumieux-Garzon et al 2018).…”

“…Third, links based on the degree of genotypic similarity of gastrointestinal microbes isolated from potentially interacting hosts within the same time frame may be used to reconstruct microbial sharing or transmission networks (VanderWaal and Ezenwa 2016;VanderWaal et al 2014). Finally, in the absence or sparsity of real data, mathematical agent-based models and artificial networks have proven exceptionally useful in modeling the transmission of parasites (Bente et al 2009;Griffin and Nunn 2012;Nunn 2009;Romano et al 2016). Such heterogeneity in connectedness within and between the components of socioecological systems may strongly influence the likelihood of parasite acquisition and transmission.…”