Files
Abstract
In Aim 1 of the study, I found heterogeneity in processing of WGS data among studies and some areas of consensus especially in recent literature. SNP thresholds are the most widely used method for inferring transmission with thresholds of 12 and 5 SNPs the most widely used. Bayesian transmission modeling attempts to address their limitation and is increasingly being used in transmission studies.
In aim 2, I investigated the role of the social network of a TB case in transmission of tuberculosis using a large social network study, the Community Health and Social Networks of TB (COHSONET) study. I also determined the relationship between genetic distance and social network distance. I found that 43% of the index case pairs who had genetically linked strains of Mycobacterium tuberculosis had an identifiable path between them in the social network, but only 13% of these index pairs were found to have a close social distance of one step in the social network. There was no correlation between genetic distance and social network distance.
In aim 3, I investigated genetic linkage among TB patients in the COHSONET study using a threshold of 12 SNPs to identify clusters of recent transmission, and covariates associated with clustering. I found that twenty-nine (36.7%) patients of the 79 sequenced isolates formed 12 clusters. A multivariate logistic analysis showed that clustered cases were more likely to be current or past smokers.
Unlike deterministic compartmental models, network models account for heterogeneity in mixing patterns. I implemented an individual-based version (particularly a network model) of a deterministic model with two latency compartments on a dynamic network simulated from a static network (Aim 4). The model depicted expected dynamics in a viability analysis when compared with a deterministic version. The model will be used to answer research questions such as whether infections in the household are sufficient to maintain the epidemic in the community, and if not so, different scenarios explaining the observed infections in the community will be simulated.
In aim 2, I investigated the role of the social network of a TB case in transmission of tuberculosis using a large social network study, the Community Health and Social Networks of TB (COHSONET) study. I also determined the relationship between genetic distance and social network distance. I found that 43% of the index case pairs who had genetically linked strains of Mycobacterium tuberculosis had an identifiable path between them in the social network, but only 13% of these index pairs were found to have a close social distance of one step in the social network. There was no correlation between genetic distance and social network distance.
In aim 3, I investigated genetic linkage among TB patients in the COHSONET study using a threshold of 12 SNPs to identify clusters of recent transmission, and covariates associated with clustering. I found that twenty-nine (36.7%) patients of the 79 sequenced isolates formed 12 clusters. A multivariate logistic analysis showed that clustered cases were more likely to be current or past smokers.
Unlike deterministic compartmental models, network models account for heterogeneity in mixing patterns. I implemented an individual-based version (particularly a network model) of a deterministic model with two latency compartments on a dynamic network simulated from a static network (Aim 4). The model depicted expected dynamics in a viability analysis when compared with a deterministic version. The model will be used to answer research questions such as whether infections in the household are sufficient to maintain the epidemic in the community, and if not so, different scenarios explaining the observed infections in the community will be simulated.