Vivli Data Challenge: Genomic-Phenotypic Association for AMR Drug Target Discovery

GEPHARD: Genomic-Phenotypic Association for AMR Drug Target Discovery Antimicrobial resistance (AMR) poses a significant threat to global health, with multidrug-resistant (MDR) pathogens becoming increasingly prevalent across sub-Saharan Africa. While phenotypic surveillance (e.g., Vivli: Pfizer – ATLAS MIC data) captures clinical resistance trends, and independent whole-genome sequencing (WGS) studies catalogue antimicrobial resistance (AMR) genes, these datasets rarely intersect, limiting their translational value for drug discovery. We hypothesize that population-level similarities in MIC resistance patterns across multiple antibiotics can be linked to specific AMR gene prevalence, allowing predictive modelling of resistance determinants and prioritization of potential drug targets. This project, GEPHARD: Genomic-Phenotypic Association for AMR Drug Target Discovery, proposes an integrative computational approach that leverages Vivli’s large-scale phenotypic antimicrobial susceptibility data and available genomic data, combined with publicly available whole-genome sequence (WGS) data from African clinical isolates, to inform drug discovery targeting multidrug-resistant pathogens. Rather than requiring isolate-level genotype-phenotype pairing, we will perform ecological association analyses, combining resistance profiles with gene prevalence data from independently sequenced genomes. This predictive gene set will form the basis for downstream target nomination, evaluating druggability, functional roles, and cross-species conservation. The project comprises three sequential objectives: (1) Surveillance Trend Analysis using Vivli’s MIC data to map MDR patterns geographically, and identify top-priority species; (2) Predictive Association integrating MIC profiles with gene prevalence to identify AMR drivers via machine learning and statistical association models; and (3) Target Nomination to prioritize AMR genes as druggable targets. This framework allows us to derive predictive associations between MIC resistance profiles and AMR gene prevalence, uncover candidate drug targets across priority African pathogens, and provide exploratory leads for compound repurposing or optimization. The outputs are anticipated to contribute to the understanding of resistance mechanisms, assist future surveillance, and guide early-stage drug discovery focused on AMR drivers prevalent in Africa.