MHCflurry documentation

MHCflurry predicts MHC class I binding affinity, antigen processing, and peptide presentation. Start with the introduction for installation and a first prediction, then choose either the command-line or Python tutorial. The reference pages are intentionally exhaustive and are best used to look up a specific option after you know which workflow you need.

Start here

• Introduction and installation explains the three prediction types and gets you to a first result.

• Command-line tutorial covers peptide prediction and protein scanning.

• Python library tutorial shows the same predictors through the Python API.

Common next steps

• Training models explains custom model fitting and release-style retraining.

• Evaluating trained models compares trained models and builds diagnostic or publication-style figures.

• Configuration and performance explains automatic hardware planning, expert overrides, and reproducibility.

• Command-line reference and API Documentation are the complete references.

Contributors and release maintainers can start with Maintainer workflows.