OluwasegunIsaac/pangea

Predictive Analytics and Genotypic Evaluation for AMR in Africa # PANGEA - Predictive Analytics and Genotypic Evaluation for AMR in Africa This README file provides information about the packages and datasets used in the code provided. # Problem Statement Antimicrobial resistance (AMR) surveillance in Africa primarily use traditional antimicrobial susceptibility testing (AST) methods, which are insufficient for early detection of pathogenic outbreaks and frequently fail to discover resistance mechanisms. Although next-generation sequencing (NGS) offers a more comprehensive understanding of these systems, its high cost and resource limitations prevent widespread adoption throughout the continent. To address these challenges, we utilised existing AST data to predict resistance gene profiles, providing a cost-effective and accessible means to gain genotypic insights. This approach enhanced the speed and accuracy of outbreak detection and bolster public health interventions in Africa, where genomic surveillance resources are limited. Our web application (PANGEA) is a web application that provides: ## Explorative Data Analysis Insightful visualisations that provide detailed explorative data summaries of the ATLAS dataset, with visualisations ranging from AMR determinant genes theme-focused to global-scale data summaries. ## Comparative Analysis In-depth comparative analyses conducted to identify predictors of poor AMR stewardship in Africa and pain points in the existing data for the surveillance work in Africa. This details a comparative markdown detailing statistical inference of data comparism, and a case study that juxtaposes the surveillance system in Africa with other countries. ## Machine Learning Models Machine Learning algorithms trained using the ATLAS dataset to predict the presence of antimicrobial resistance genes in bacterial samples and also predict the resistance status based on a complex network of sample metadata and specific subsidiary data as input parameters. The models' performance, accuracy, and efficiency a …