Development of ABR package using machine learning and AI tools

Compilation of data base using real time sequencing data and development of anti bacterial resistance (ABR) package using machine learning tools for sustainable solutions towards emerging ABR challenges faced by low middle income countries (LMICs). It would be product based solution for the pharmaceuticals, hospital, pathogen laboratories and research institutions.

Dr. Imran Hashmi is the Primary Investigator, serving as Tenured Professor and Associate Dean at School of Civil and Environmental Engineering, Institute of Environmental Sciences and Engineering, NUST.

The development and spread of ABR is a community phenomenon that links reservoirs across multiple scales both in the environment and within the macrobiotic. Common sources responsible for the presence of pharmaceutical compounds in the environment are hospitals, municipal wastewater, livestock farms and pharmaceutical manufacturing units. As a result, receiving water bodies like rivers, dams etc. become contaminated with antibiotics and the compounds become part of the food chain and causing alarming risk of ABR as pathogens becoming more resistant toward modern medicines. There are 4.95 million deaths are reported per year due to antimicrobial resistance as per WHO 2023. Situation is worse in Low- and middle-income countries as local communities bear high burden of drug resistant infectious disease. Antimicrobial resistance occurs when pathogens including bacteria, fungi, viruses and parasites evolve over time and become resistant to existing treatment methods was described by WHO (2019) as one of the serious threat to the public health globally. Its a major health crisis faced by humans therefore, a sustainable solution is necessary to mitigate this challenge. The proposed solution will address ABR issues by studying the bacteria spread and their subsequent effects in detail using modern machine learning and Artificial intelligence technologies.

The target audience are hospital treatment facilities, research centers and the pharmaceutical companies. The ABR being the most neglected topic in the LMIC with a detrimental impact on the human beings and animals has been studied by few research institutes. Therefore, a provision of a comprehensive database to address such a challenge is the dire need of the stakeholders. As this study aims to develop such package it can provide a critical support to them. This need has been understood by overall surveying the ABR scenario in the LMIC and how the antibiotics are getting futile in patients treatment. Therefore, the stakeholders needs a compact database/package to get an insight on how the ABR is evolving and what counter measures; in what direction are required. The stakeholders are being engaged by reaching out to them through interviews and surveys with the medical doctors and the pharmaceutical R&D departments that what difficulties are they facing while improving the new antibiotic and assessing the situation and making them realize that how crucial is the formulation of a compact ABR package to devise a permanent solution to address ABR challenge sustainably.

This project is fundamentally interdisciplinary and it has the potential to benefit all these disciplines by introducing new ideas and new ways of working together. Specifically:

• Environmental genomics: effective study design is the key to ensuring robust and representative findings from metagenomics surveys.
• Machine learning: A comprehensive data base established on genomics and sequencing result will be helpful to hospital treatment facilities to recommend effective antibiotics to end user to mitigate resistant pathogenic strains.
• Health systems research: typically, these studies do not have direct access to actual biological data on prevalence and transmission of AMR genes. The study will illustrate how to bring a new quantitative dimension to these types of analyses through the incorporation of such data.
• Epidemiological modelling: this will be the first study to parameterize epidemiological models of ABR gene spread. This will create a major impact in the field and possibly lead to multiple novel avenues of modelling.
• One Health (OH): engagement of mathematical modelers with researchers across the OH spectrum will allow a comparison of the impact of alternative drivers of AMR (for example waste from pharmaceutical companies verses large cattle farms), which has never been done before in LMIC contexts. 

The resulting data will be used to parameterize predictive models of the spread of the ABR in health facilities and the environment. These will be used to evaluate the most socially and politically feasible intervention strategies as identified by concurrent health policy and systems research. This will provide crucial information to policy makers and stakeholders to pursue policies that impact ABR with consequent benefits to the local community.

Target population:

The solution will address LMICs specifically. This research will impact on the policies and responses for tackling drivers of ABR being developed in LMICs. It is obviously relevant to Pakistan and it will have a regional impact too. Furthermore, focusing  AMR challenges associated with extreme urbanization, a common problem facing LMICs, will give it global relevance. It will also be amongst the first studies analyzing the policy process relating to AMR in LMICs, which is little understood and will be critical to implementation. These outputs will be applicable beyond Pakistan. The factors driving ABR in Pakistan, including intensive agriculture and inappropriate usage of antibiotics, are common to many LMICs; our findings and methods developed will therefore have a wider relevance particularly to other South Asian countries such as India. 

This project has a substantial, and very well-integrated, social science element. Therefore, the project is an ideal vehicle with which to inform the public of AMR issues. End-users would be engaged in identifying drivers of AMR in the Pakistan context through a systems-mapping exercise. This approach will be further extended to use our materials and results to create educational videos to be made public on the internet. Social media platforms will be used in this activity as this will include retweets of the Twitter account specifically curated for this project, where information on findings and progress will be disseminated to the wider community. The mathematical and statistical models will be explicitly designed to allow their application to other countries following suitable reparameterization and we will identify additional data sources for doing this during the analysis phase of the project. While the specific policy options will be tailored to the Pakistani context, critical research questions about stakeholder motivations and support for or opposition to alternative policy options are relevant widely. 

Project will be divided in different phases according to research objectives and progress will be evaluated by setting parameters with the help of defined timeline.

1. Academic outputs including scientific papers, conference presentations, epidemiological models and policy decision-making frameworks for research will be designed with both international, national, and regional academic institutes.
2. ABR package that will help pharmaceuticals, hospitals and research institutions.

In addition to measure overall success following indicators will help.

• Reduction in Incidence of Resistant Infections: The primary goal of ABR initiatives is often to reduce the incidence of infections caused by antibiotic-resistant bacteria. Success can be measured by tracking trends in the prevalence of resistant infections over time, comparing rates before and after intervention implementation.
• Antibiotic Consumption Patterns:Monitoring antibiotic consumption patterns is critical for assessing the impact of interventions on antibiotic use. Success may be measured by evaluating changes in antibiotic prescribing practices, such as a decrease in inappropriate antibiotic prescriptions or a shift towards narrower spectrum antibiotics.
• Resistance Trends: Monitoring resistance trends in targeted pathogens provides valuable insights into the effectiveness of interventions. Success can be measured by observing decreases in the prevalence or incidence of specific resistant strains in clinical and community settings.

In order to effectively beat the threat of antibiotic resistance (ABR) posed to global health, it is is crucial to develop effective strategies to combat ABR. In the context of Pakistan, there are several barriers associated with the high risk of AMR. 

Some of the barriers are explained following.

• Shortage of skilled human resource regarding ABR related issues: Lack of trained medical practitioners with expertise in antibacterial resistance related infectious diseases is one of the barrier in mitigating ABR.
• Poor public awareness regarding ABR related problems: As ABR is a neglected issues in Pakistan, there is knowledge gap in public as well as in institutions. Lack of public awareness leads to self medication practices, over consumption and misuse of antibiotics without prescriptions ultimately contributing in high ABR risks. 
• Inadequate ABR surveillance infrastructure: No surveillance structure is present in Pakistan to monitor antibacterial resistance. There are no such initiatives to assess and mitigate ABR challenges.
• Lack of ABR advocates for surveillance & research: Due to lack of surveillance no such policies exist to investigate and combat growing problem of ABR. 

It would help in training people in terms of capacity building to overcome ABR problems, constructing ABR surveillance infrastructure and research programs to investigate route causes, intensity of problem and reliable solutions. Trinity challenge can help in providing expertise to execute training programs aimed on capacity building and skilled human resource. It will empower healthcare professionals, environmental researchers and govt officials (policy makers) to effectively mitigate ABR challenge. Development of strong surveillance infrastructure is essential for the monitoring the prevalence of ABR and its emerging challenges. The trinity challenge can support the infrastructure development through providing funds for the ABR research program, implementation of advanced research methods, and upgrading research instrumentation. Furthermore, proposed solution is seeking funding to overcome the financial barrier to carry out comprehensive research about antibacterial resistance. Summarizing, The Trinity Challenge can play a crucial role to overcome barriers in addressing ABR by strengthening capacity building, developing surveillance infrastructure, initiating research programs, and funding ABR research and intervention. By investing in such areas this challenge can introduce collaborative efforts to combat ABR issues.

The partners include the following.

• Dr. Christopher Quince, MRC Principal Research Fellow, Warwick Medical School, University of Warwick. Dr. Chris is Professor at Earlham Institute UK. He is an expert in microbial bioinformatics and multivariate statistics. Dr. Chris Quince has pioneered techniques to reduce sequencing noise and the application of methods such as shotgun metagenomics to microbial communities. This cutting-edge approach fragments and sequences all the genomes in a microbiome quite literally taking a shotgun approach and these are then reconstructed using sophisticated computational approaches. The methods to reconstruct sequence fragments back into metagenome assembled genomes (MAGs) have been significantly advanced by Quince. He has used these metagenome techniques and analysis methods, in conjunction with his wide network of collaborators, to address urgent clinical questions, such as trialing new treatments for pediatric Crohn’s disease and understanding what drives antimicrobial resistance (AMR) in the gut microbiome. He is commencing the Global Challenges Research Fund GastroPak project, which aims to determine the sources of non-viral gastroenteritis in Pakistan.