Animal tuberculosis (TB), primarily caused by *Mycobacterium bovis*, poses a significant threat to livestock, wildlife, and human health globally. Portugal grapples with *M. bovis* circulating in a multi-host system involving cattle, red deer, and wild boar. While interspecies transmission has been documented, the ecological drivers of transmission remain unclear. This study aims to reconstruct the spatiotemporal history of animal TB in Portugal and to clarify the *M. bovis* population structure to inform improved control strategies. The persistence of *M. bovis* across multiple host species complicates eradication efforts. Cattle are considered the primary host, but various wildlife species can contribute to pathogen maintenance, depending on the ecosystem and epidemiological context. Molecular studies have demonstrated interspecies transmission between cattle and wildlife, with wildlife implicated as a risk factor for cattle TB infection. In the Iberian Peninsula, red deer and wild boar are significant wildlife reservoirs. Portugal has implemented a national eradication program since 1987, with significant progress made. However, regional variations in TB prevalence persist, with hotspots in central and southern regions where livestock and wildlife extensively interact. The high host densities and extensive management practices in these areas are suspected to promote transmission. To address this complex situation, the researchers focused on high-burden areas with significant livestock-wildlife interaction, utilizing a large collection of *M. bovis* isolates gathered over nearly two decades.

The introduction section provides a detailed review of existing literature on *Mycobacterium bovis*, its multi-host dynamics, and the challenges of TB eradication. It cites several studies demonstrating interspecies transmission between cattle and various wildlife hosts, such as badgers in the UK and Ireland, African buffalo in South Africa, brushtail possums in New Zealand, and red deer and wild boar in the Iberian Peninsula. The review also discusses the challenges in controlling TB in wildlife populations and the national eradication program in Portugal, highlighting the regional variations in TB prevalence. Several studies are referenced that detail the occurrence of *M. bovis* in wild populations in Portugal and Spain and the growing concern over the abundance of wild ungulates and the potential for disease expansion.

The study employed a comprehensive methodology involving molecular characterization and statistical analyses of 948 *M. bovis* isolates collected between 2002 and 2016 from cattle, red deer, and wild boar in three major TB hotspot regions in Portugal (Castelo Branco, Portalegre, and Beja). Isolates were characterized using spoligotyping (n=948) and MIRU-VNTR (n=524). Spoligotyping identified 64 distinct profiles, while MIRU-VNTR distinguished subtypes within each spoligotype. The discriminatory power of both methods was calculated. A representative subset of isolates (n=487) was subjected to more detailed analyses, including Bayesian population structure inference using STRUCTURE software, Minimum Spanning Tree (MST) analysis using BioNumerics, and network analyses using Gephi. A space-time cluster analysis using SaTScan identified statistically significant clusters of TB cases, assessing the influence of host species on spatiotemporal risk. Data on herd or hunting area origin was used to identify small outbreaks versus chronically infected scenarios. The study also analyzed the allelic richness of *M. bovis* populations using HP-RARE software and used Chi-squared tests to assess the statistical significance of associations between spoligotyping profiles/ancestral populations and geographic regions or host species. The dataset was split into three temporal periods (2002–2008, 2009–2012, 2013–2016) to reflect changes in surveillance intensity and TB prevalence.

Spoligotyping revealed 64 profiles, with six accounting for 65% of isolates. MIRU-VNTR provided finer discrimination (D=0.97 vs 0.92 for spoligotyping). Bayesian analysis identified five ancestral *M. bovis* populations (AM1-AM5). AM4 showed a strong association with cattle and the Beja district, suggesting a cattle origin for the oldest population. Spatiotemporal analysis identified two significant clusters: one in Beja (2004-2010), dominated by cattle; and one in Castelo Branco (2012-2016), dominated by wild boar. Network analysis revealed strong regional connections among host species, especially within Castelo Branco and between Castelo Branco and Portalegre. Analysis of chronically infected herds/hunting areas showed higher diversity in wildlife compared to cattle. MST analysis grouped most isolates into 14 clonal complexes, with some showing host or geographic specificity. The most prevalent spoligotypes were found in both cattle and wildlife, highlighting cross-species transmission. Analyses of allelic richness suggest that AM4 holds the highest mean allelic richness, indicating a possible older origin for this population. This was further corroborated through MST analyses.

The findings confirm the complex interplay of livestock and wildlife in maintaining and transmitting *M. bovis*. The strong association of the oldest population (AM4) with cattle in Beja suggests a likely cattle origin for the TB epidemics. The two distinct spatiotemporal clusters highlight the changing roles of different hosts over time and across regions, reflecting the impact of surveillance and control measures. The high genotypic diversity in wildlife, especially in chronically infected hunting areas, emphasizes their significance as reservoirs. The high degree of overlap in genotypes found in different host species strengthens the hypothesis of both direct and indirect transmission routes, linking cattle and wildlife. The study's findings are consistent with other studies demonstrating geographic structuring of *M. bovis* and the impact of wildlife behavior and density on transmission risk. The limitations in wildlife sampling and potential homoplasy in the typing methods were acknowledged, impacting on the precision of the transmission events inferred. The data suggest that the ongoing evolution of the *M. bovis* population necessitates continued surveillance and adaptation of control strategies.

This long-term molecular epidemiological study provides crucial insights into the dynamics of *M. bovis* in Portugal. The data strongly suggest a cattle origin for the TB epidemics, with subsequent spread to wildlife populations. The identified spatiotemporal clusters, genotypic diversity patterns, and host-specific associations highlight the need for a holistic control strategy that addresses both livestock and wildlife reservoirs. Future research employing whole-genome sequencing could provide even more detailed insights into transmission patterns and microevolutionary events within *M. bovis* populations.

The study acknowledges several limitations. Wildlife sampling was not consistently implemented throughout the 15-year study period. Thus, the observed patterns for wildlife hosts may be influenced by this sampling bias. Although a combined approach using spoligotyping and MIRU-VNTR aimed to minimize the impact of homoplasy, the possibility of homoplasy could still affect the accuracy of transmission inference. Furthermore, the study focuses on a limited number of molecular markers; whole genome sequencing might reveal additional information about the evolutionary dynamics and transmission of the pathogen. The analysis doesn't fully explore the environmental contamination role as a transmission route.