Machine Learning Approaches for Predicting Antimicrobial Resistance (AMR)
Predicting Tomorrow’s Drug Resistance with Today’s Machine Learning
Skills you will gain:
About Program:
Antimicrobial resistance is a growing global threat, driven by rapid evolution of microbes and increased horizontal gene transfer. Traditional laboratory-based AMR testing is slow and resource-intensive, creating a need for computational tools that can rapidly predict resistance patterns. Machine learning provides powerful approaches to detect resistance markers, classify microbial strains, and predict phenotypic resistance using genomic and metagenomic features.
This workshop blends microbiology and AI, teaching participants how to preprocess AMR datasets, extract meaningful features, train ML models, and validate predictive performance. Real-world case studies will highlight applications of ML in clinical diagnostics, public health surveillance, environmental AMR monitoring, and drug discovery. Participants will learn to integrate ML pipelines with well-known AMR databases and interpret model outputs for actionable insights.
Aim: This workshop aims to introduce participants to AI- and machine-learning–based techniques for predicting antimicrobial resistance (AMR) from genomic, metagenomic, and clinical datasets. It provides foundational and advanced understanding of resistome profiling, feature engineering, and predictive modeling of resistant phenotypes. Hands-on sessions will equip learners with practical skills in building, evaluating, and interpreting ML models for AMR surveillance and research. The workshop bridges computational methods with real-world AMR challenges in healthcare and environmental microbiology.
Program Objectives:
- Learn AMR mechanisms and genomic markers associated with resistance.
- Apply ML algorithms (RF, SVM, XGBoost, deep learning) for resistance prediction.
- Perform data preprocessing, feature extraction, and model building on AMR datasets.
- Use major AMR databases (CARD, ResFinder, MEGARes) for annotation.
- Evaluate predictive performance using precision, recall, F1, AUROC, and interpretability tools.
What you will learn?
Day 1 — Foundations of AI in AMR
- Lecture: Overview of AMR mechanisms and AI’s role in prediction.
- Session: Understanding genomic and phenotypic data formats.
- Hands-on: Data preprocessing — cleaning, encoding categorical data, normalization.
- Case Study: Exploring public AMR datasets (PATRIC, CARD, MEGARes).
Day 2 — Feature Engineering and Model Training
- Lecture: Feature extraction — k-mers, SNPs, and gene presence matrices.
- Hands-on: Building ML pipelines with scikit-learn.
- Exercise: Training Random Forest, SVM, and XGBoost models.
- Activity: Hyperparameter tuning and model comparison.
Day 3 — Evaluation, Interpretation & Visualization
- Hands-on: Model evaluation — confusion matrix, ROC-AUC, precision–recall.
- Lecture: Model interpretability — feature importance, SHAP values.
- Case Study: Predicting multidrug resistance patterns.
- Practical: Visualization of prediction outputs using Python.
- Discussion: Integration of ML outputs into AMR monitoring systems.
Mentor Profile
Fee Plan
Get an e-Certificate of Participation!
Intended For :
- Undergraduate/Postgraduate degree in Microbiology, Biotechnology, Bioinformatics, Computational Biology, or related fields.
- Professionals in clinical microbiology, healthcare, pharmaceutical R&D, or public health sectors.
- AI/ML engineers & data scientists interested in applying models to biological and public health problems.
- Individuals passionate about AMR prediction, microbial genomics, and data-driven epidemiology.
Career Supporting Skills
Program Outcomes
- Ability to build ML models for predicting antimicrobial resistance.
- Competence in using genomic and metagenomic features for AMR prediction.
- Experience with AMR databases for annotation and feature extraction.
- Practical knowledge of ML evaluation and performance metrics.
- Skill in interpreting ML results for biological and clinical relevance.