Integrative bioinformatics and machine learning approaches to investigate the emergence and transmission of antimicrobial resistance

Abstract

Antimicrobial resistance (AMR) represents a major global health challenge affecting human, animal, and environmental systems. Advances in genomic and metagenomic sequencing have greatly improved our ability to characterize resistance determinants. The purpose of this thesis was to investigate AMR patterns by integrating genome-based analyses, metagenomic data, and computational approaches, including bioinformatics and machine learning (ML), within a One Health framework, which provided complementary perspectives on AMR surveillance, resistance gene diversity, and the ecological and evolutionary contexts that shape contemporary resistance patterns.

In Chapter 3, we analyzed a large, longitudinal collection of multidrug resistant Salmonella enterica isolates from multiple food-chain sources in China to investigate the genomic determinants underlying resistance dissemination, persistence, and future trajectories. By integrating whole genome sequencing with ML and predictive modelling, this chapter identified clinically relevant, mobile resistance traits that are widely distributed across human, food, and agricultural reservoirs. Food workers emerged as key transmission nodes, showing higher trait prevalence than food sources, and uniquely carrying last-resort antibiotic resistance genes. Finally, the results demonstrated that socioeconomic conditions and patterns of antimicrobial consumption are important drivers shaping the future expansion of AMR within food production systems.

In Chapter 4, this thesis examined global patterns of antimicrobial resistance by integrating large-scale genomic data with antimicrobial susceptibility profiles and a wide range of social, economic, and environmental indicators. Using a ML–based, multi-scale analytical framework, this chapter identified pathogen-specific resistance-associated genomic traits with projected increasing trends and assessed the broader contextual factors linked to their emergence and expansion. The results highlight the strong association between AMR dynamics and socioeconomic conditions, demonstrating how disparities in social and environmental contexts can shape global resistance trajectories.

In Chapter 5, ancient host-associated metagenomes spanning hundreds to tens of thousands of years were analyzed to investigate AMR in a long-term historical context. Despite challenges related to DNA degradation and fragmentation, metagenomic and phylogenetic analyses enabled reconstruction of taxonomic context and screening of putative antimicrobial resistance gene, providing insights into the deep history and diversity of resistance-related genetic elements.

Taken together, this thesis demonstrates that integrating genomic and metagenomic data with bioinformatic and ML–based analyses across contemporary and historical timescales can improve understanding of AMR patterns and support the identification of resistance determinants with potential future impact. This work contributes to the development of more informative AMR surveillance strategies within a One Health framework.

Date of Award	18 Jul 2026
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Tania Dottorini (Supervisor) & Weihua Meng (Supervisor)