Contemporary patterns of land use and global climate change are modifying regional pools of parasite host species. The impact of host community changes on human disease risk, however, is difficult to assess due to a lack of information about zoonotic parasite host assemblages. We have used a recently developed method to infer parasite-host interactions for Chagas Disease (CD) from vector-host co-occurrence networks. Vector-host networks were constructed to analyze topological characteristics of the network and ecological traits of species´ nodes, which could provide information regarding parasite regional dispersal in Mexico. Twenty-eight triatomine species (vectors) and 396 mammal species (potential hosts) were included using a data-mining approach to develop models to infer most-likely interactions. The final network contained 1576 links which were analyzed to calculate centrality, connectivity, and modularity. The model predicted links of independently registered Trypanosoma cruzi hosts, which correlated with the degree of parasite-vector co-occurrence. Wiring patterns differed according to node location, while edge density was greater in Neotropical as compared to Nearctic regions. Vectors with greatest public health importance (i.e. Triatoma dimidiata, T. barberi, T. pallidipennis, T. longipennis, etc), did not have stronger links with particular host species, although they had a greater frequency of significant links. In contrast, hosts classified as important based on network properties were synanthropic mammals. The latter were the most common parasite hosts and are likely bridge species between these communities, thereby integrating meta-community scenarios beneficial for long-range parasite dispersal. This was particularly true for rodents, >50% of species are synanthropic and more than 20%