immuneML

immuneML is a machine learning framework for the analysis and classification of adaptive immune receptors and repertoires (AIRR), developed by ELIXIR Norway and initially described in Pavlovic et al. It is designed for researchers working with immune receptor data who need to train ML models, apply them to new datasets, simulate benchmarking data, and perform exploratory analyses.

The framework supports research across a range of applications including immunotherapy, vaccine development, and disease diagnostics, with workflows defined through YAML-based specifications to ensure reproducibility.

Key Features

• Repertoire classification — train models for disease prediction based on immune repertoire data
• Receptor sequence classification — predict antigen binding and other receptor-level properties
• Dataset simulation — generate synthetic AIRR datasets for ML model benchmarking
• Exploratory analysis — statistical analyses, dimensionality reduction, and clustering for deeper insight into immune data
• Reproducible workflows — YAML-based specification of complete analysis pipelines
• Broad applications — supports research in immunotherapy, vaccine development, and disease diagnostics

Getting Started

1. Access immuneML through the web interface
2. Follow the Quickstart tutorial to set up a first analysis
3. Define the analysis using YAML specifications or use provided templates
4. Train, evaluate, and apply models to AIRR datasets

immuneML is accessible via a web platform that also hosts documentation and templates. The framework is provided by ELIXIR Norway as part of its suite of bioinformatics services.