A novel AI laboratory platform to identify bacteria and AMR data challenge

Antimicrobial resistance (AMR) is considered a global public health threat by several health-focused, research organizations and academic institutions including the World Health Organization and similar sister organization like the Food and Agriculture Organization of the United Nation. For a very long time, Hospital AMR surveillance is regarded as an essential strategy to identify resistant bacteria, their spread and prevalence. On the contrary, this strategy of AMR surveillance developed around culture-based isolation of bacteria from hospital patients, and their phenotypic response to a battery of antibiotics has been shown to be less representative and time-consuming. These time-consuming techniques coupled with poor laboratory information management systems seen within several clinical and research laboratories in Ghana and several resource limited environments have contributed to the lack of data on pathogens of public health importance and their epidemiology. Wastewater-based surveillance is increasingly being recognized as an important approach to monitoring population-level AMR. Therefore, we hypothesized and proposed a pilot project to decipher the nexus of AMR distribution using digital innovation and molecular detection using wastewater samples in Ghana. From the results of this pilot using 171 wastewater samples, we identified 45% (77/171) E. coli contaminated wastewater with 62% of them showing ESBL resistance. As part of this data challenge we propose to answer the following questions; 1) whether or not we can compare results from our pilot study to extended hospital and community based surveillance datasets from big pharma. 2) Understand AMR trends for E. coli, its associated mechanisms comparing our genotype data to those collected in the data challenge database we requesting. 3) Validate our AMR detection and bacteria identification platform using the requested database.