Using genomic data and machine learning to predict antibiotic resistance: A tutorial paper.

Area of Science:

• Microbiology
• Bioinformatics
• Computational Biology

Background:

• Antibiotic resistance is a significant global health threat, necessitating advanced diagnostic tools.
• Genomic sequencing combined with machine learning (ML) offers potential for improved accuracy and cost-effectiveness in detecting bacterial drug resistance.

Purpose of the Study:

• To provide a beginner-friendly, step-by-step tutorial for training and evaluating ML models to predict antibiotic resistance.
• To equip pre-health and life sciences students with practical skills in applying ML to medical challenges.

Main Methods:

• The tutorial guides users through data preparation and training of four ML models: logistic regression, random forests, extreme gradient-boosted trees, and neural networks.
• Model performance is evaluated using various metrics and cross-validation techniques.
• The tutorial is implemented in Google Colab notebooks, requiring no software installation.

Main Results:

• The tutorial successfully demonstrates the process of building and assessing ML models for predicting drug resistance in Escherichia coli.
• It provides a foundational understanding of different ML algorithms and their application in microbiology.

Conclusions:

• Machine learning presents a valuable approach for enhancing antibiotic resistance testing.
• This tutorial serves as an accessible educational resource for students to learn essential ML skills for future medical applications.