From sample to insight: Onsite training and implementing mobile environmental surveillance in sub-Saharan Africa

Environmental surveillance for human pathogen detection is crucial for safeguarding public health; however, traditional approaches are often time-consuming, equipment-intensive and workflow-complex, delaying actionable public-health decision-making. The integration of a mobile laboratory (ML) with rapid bioinformatic analysis and interactive data visualization offers a novel approach to this surveillance. The ML workflows, with a focus on metagenomic sequencing and qPCR technologies, have been specifically designed to provide real-time information in detecting epidemiological threats in environmental samples. We have adopted a minimalist strategy, avoiding complex, expensive, and energy-intensive equipment that is challenging to maintain and transport, thereby reflecting a more sustainable approach. This study details the activities and outcomes of implementing such a ML in Tanzania. During one-month ML deployment, a total of 83 samples were sequenced using long-read metagenomic sequencing, and 70 samples were analyzed by Biomeme qPCR. The deployment enabled the detection of multiple bacterial, viral, and antimicrobial resistance targets in wastewater, drinking water, surface water, and soil samples providing valuable insights into the population epidemiological information relevant to the prevention of waterborne and zoonotic infections. In addition, the ML deployment provided a rare opportunity for knowledge transfer, actively engaged local researchers, and strengthened local epidemiological capacity. After the deployment, certificates of completion were issued to twenty-five scientists from four African countries (Tanzania, Burkina Faso, Democratic Republic of Congo, and Ethiopia) and post-deployment surveys were sent out to gather their feedback. Our findings highlight the importance of combining targeted qPCR assays with metagenomic sequencing for the proper interpretation, while acknowledging the risk of false-positive detections arising from low-abundance sequencing reads, particularly when using fast taxonomic classification tools such as kraken2. Given the hardware limitations inherent to ML deployments, metagenomic pathogen detection in this context should therefore be regarded primarily as a valuable screening tool rather than a definitive diagnostic method. These observations further underscore the need for more stringent and computationally demanding bioinformatic approaches when assessing high-priority pathogens. Complementary qPCR-based assays conducted within the ML can ensure equivalent robustness to those conducted in conventional laboratory settings. This innovative approach has the potential to significantly enhance detection capabilities and support timely public health responses in resource-limited settings.