• Corpus ID: 251719203

Spatial modeling of dyadic genetic relatedness data: Identifying factors associated with M. tuberculosis transmission in Moldova

@inproceedings{Warren2021SpatialMO,
  title={Spatial modeling of dyadic genetic relatedness data: Identifying factors associated with M. tuberculosis transmission in Moldova},
  author={Joshua L. Warren and Melanie H. Chitwood and Benjamin Sobkowiak and Valeriu Crudu and Caroline Colijn and Ted Cohen},
  year={2021}
}

• J. Warren, M. Chitwood, +3 authors T. Cohen
• Published 28 September 2021
• Psychology

Understanding factors that contribute to the increased likelihood of disease transmission between two individuals is important for infection control. However, analyzing measures of genetic relatedness is complicated due to correlation arising from the presence of the same individual across multiple dyadic outcomes, potential spatial correlation caused by unmeasured transmission dynamics, and the distinctive distributional characteristics of some of the outcomes. We develop two novel hierarchical Bayesian spatial methods for analyzing dyadic genetic relatedness data, in the form of patristic distances and transmission probabilities, that simultaneously address each of these complications. Using individual-level spatially correlated random effect parameters, we account for multiple sources of correlation between the outcomes as well as other important features of their distribution. Through simulation, we show the limitations of existing approaches in terms of estimating key associations of interest, and the ability of the new methodology to correct for these issues across datasets with different levels of correlation. All methods are applied to Mycobacterium tuberculosis data from the Republic of Moldova where we identify previously unknown factors associated with disease transmission and, through analysis of the random effect parameters, key individuals and areas with increased transmission activity. Model comparisons show the importance of the new methodology in this setting. The methods are implemented in the R package GenePair . We develop models to analyze spatially-referenced dyadic genetic relatedness outcomes (i.e., patristic distances, transmission probabilities) while accounting for multiple sources of correlation between responses and important features of the outcomes. The methods are designed to yield accurate statistical inference for the primary regression associations of interest while also identify-6 previously observed) and then collect samples from the posterior predictive distribution of the random effect parameters at each location separately (e.g., samples from f t ln P . . . , ln the patristic distances model) using techniques from ? . These samples are then summarized using posterior means and standard deviations and plotted across the domain using inverse distance weighting to fill in the remaining gaps (for visualization purposes only). Areas in red from the posterior mean map suggest that patristic distances/transmission probabilities involving individuals residing in these locations tend to be smaller/larger on average than areas in blue, possibly indi-cating geographic areas of increased residual transmission activity. For patristic distances however, the variability in these values overall is not large in this setting, suggesting that the risk does not differ substantially across the map.