AI-driven rapid, high-capacity AMR surveillance in environment

Samples from drinking water purification plant inlets will be screened using a rapid colorimetric assay which autonomously identifies positive samples, and drives an AI-powered satellite image analysis to identify sources and risk factors of AMR to develop a "catchment pollution index", with ground truthing done by whole genome sequencing. 

Prof. Rasika Jinadasa, BVSc (Peradeniya), MS (Nebraska), PhD (Cornell) Professor in Microbiology, Department of Veterinary Pathobiology

And

Coordinator, Postgraduate Education Unit,  Faculty of Veterinary Medicine & Animal Science

University of Peradeniya, 

Peradeniya, Sri Lanka.

Sri Lanka lacks a surveillance system for AMR in the environment, therefore the true burden is unknown, and no effective remedial measures are practiced. Few recent publications and our preliminary work show that major rivers in Sri Lanka are contaminated with AMR bacteria (ARB), and human activity clearly increases ARB in water bodies. These rivers provide drinking water to most of the population, and irrigation for most of the leafy green vegetable farms, putting the consumers at risk. These ARBs frequently originate from farm effluent, and improperly managed sewage systems. Our work shows that vegetable farms that use poultry manure are significantly more contaminated with ARBs than farms that do not use it, and the ARBs may prevail in the farm soil long after the secession of the manure use. Similarly, we have shown that soil in dairy farms contain multidrug resistant (MDR) ARBs. Our work also shows that ARB burden in major rivers changes with the rainfall patterns, and it does not necessarily correlate with the population density of the catchment areas. Current culture dependent antimicrobial resistance detection methods are time consuming (>5 days), not cost effective and can not be used for high throughput screening of AMR.

Large proportion of the population in Sri Lanka who are the water consumers long the two largest rivers in the country will be benefitted. Similarly, >50% of the urban population that consume the leafy green vegetables (who may consume produce grown in fields irrigated by these rivers) will also be benefitted.  The solution identifies and prioritizes preventable sources of AMR contaminations in the environment (specially in water sources), and therefore, in a larger scale, the entire population will be benefitted. eal-time data accessibility is facilitated through a dedicated website, with plans for a mobile app enabling public viewing of water quality data and inputting antibiotic usage details, supported by AI-driven interpretation.

The solution identifies and prioritizes preventable sources of AMR contamination in the environment, that may effectively mobilize limited funding sources available in LMCs such as Sri Lanka. It is proposed to develop a policy brief to the government of Sri Lanka (white paper) and publish all findings in peer-reviewed journals, and disseminated through regular mass media. There will also be a website for real time reporting and the entire technological know how will be open-source model resulting in a free-to-use dashboard.

This will initially benefit a large population (water consumers along the two largest river systems, and a large portion of vegetable consumers). in the long run it will benefit more people as  policy decisions that may follow will address the remedial measures for preventable sources of AMR pollutants.  

The resultant technology can easily be used across the entire hydro catchment areas of the island, and thereby monitor a a significantly large part of the country. It can be easily adapted to  regional countries with similar geoclimatic backgrounds. 

The reduction of AMR in the environment will be visible on the  data at the screening level. One remedial measures are adopted in a a given catchment are, AI-driven monitoring can evaluate the changes in the land use patterns and provide further predictions for improvement.  

Depreciation of the local currency following the economic crisis and reduced imports, in the country has dramatically increased the cost of high-end computers required for the project. To mitigate this issue, remote access/cloud computing may be sourced during the early stages of the project. 

As the economy is improving and the inflation has slowed down (5.9% at present), the costs are expected to reduce by next year. 

Food and Agriculture Organization of (FAO)

World Health Organization (WHO)

World Bank 

United Nations Environmental Programme (UNEP)

International Centre for Antimicrobial Resistance Solutions (ICARS)

Technical University of Denmark 

Quadripartite AMR multi stakeholder platform