How AI is helping scientists track viruses through wastewater

Wastewater surveillance has long tracked pathogens such as poliovirus and typhoid, but AI is transforming the process by enabling high-throughput metagenomic sequencing. This method captures genetic fragments from all organisms in a sample, allowing detection of unexpected or emerging viruses without relying on clinical reports. Recent applications identified a single measles case in Cook County, Illinois, across a catchment of over one million residents, demonstrating the sensitivity of the approach.

Traditional laboratory methods often struggle with low-frequency viral detection in environmental samples, while metagenomic sequencing produces vast, noisy datasets that require extensive computational processing. Microsoft’s Premonition Insights addresses this bottleneck by using massively parallel workflows on Microsoft Azure to compare raw sequences against comprehensive organism databases, resolving ambiguous alignments and filtering noise with advanced statistical models.

Even after cloud-based processing, researchers face a secondary challenge: transforming terabytes of sequencing data into actionable epidemiological insights. The Microsoft Discovery app streamlines this phase by enabling domain experts to orchestrate AI agents for data preparation, visualization, and hypothesis generation, while maintaining human oversight for scientific interpretation and quality control.

In a case study using publicly available wastewater data from Gujarat, India, researchers applied the Discovery app to analyze over 500 samples collected across four cities over a year. The workflow revealed seasonal patterns in enteric viruses like rotavirus and adenovirus, with peaks aligning with the monsoon season, and identified clusters of enterovirus sequences that may indicate local outbreaks, though no poliovirus was detected.