LlamaCancer

LlamaCancer is a comprehensive Python framework for analyzing biomarker associations in oncology clinical trials. It provides a streamlined workflow for biomarker analysis, including data loading, processing, statistical analysis, and visualization.

Features

• Flexible data loading: Support for various data formats and structures
• Automated biomarker dichotomization: Convert continuous biomarkers to categorical (high/low) groups using multiple methods
• Comprehensive statistical analysis: Log-rank tests, Cox proportional hazards models, Fisher's exact tests, and more
• Publication-quality visualizations: Kaplan-Meier plots, forest plots, volcano plots, and more
• Configuration-based workflow: Define analysis parameters in reusable configuration files
• Extensive documentation: Comprehensive user guide, API reference, and tutorials

Installation

# Install from PyPI
pip install llamacancer

# Install from source
git clone https://github.com/llamagroup/llamacancer.git
cd llamacancer
pip install -e .

Quick Start

import llamacancer as lc
from llamacancer.config import load_config
from llamacancer.io import load_clinical_data, load_biomarker_data, merge_clinical_biomarkers
from llamacancer.analysis import run_biomarker_associations

# Load configuration
config = load_config("configs/default_analysis_config.py")

# Load and merge data
clinical_df = load_clinical_data(config)
biomarker_df = load_biomarker_data(config)
merged_df = merge_clinical_biomarkers(clinical_df, biomarker_df)

# Run biomarker association analysis
results = run_biomarker_associations(merged_df, config)

# Display significant biomarkers
print(f"Significant biomarkers: {results['summary']['significant_biomarkers']}")

Example Workflow

1. Define your configuration:

   # configs/my_analysis_config.py
   from ml_collections import config_dict
   
   def get_config():
       config = config_dict.ConfigDict()
       config.project_name = "My Biomarker Analysis"
       config.data_dir = "data/"
       config.biomarkers_to_analyze = ["B_cell_GES", "CD19_Expression_Level"]
       # ... more configuration options
       return config
   

2. Prepare your data:

   • Clinical data CSV with patient identifiers, treatment arms, endpoints
   • Biomarker data CSV with patient identifiers and biomarker measurements

3. Run the analysis from command line:

   llamacancer --config configs/my_analysis_config.py
   

4. Or run interactively in a notebook:

   jupyter notebook notebooks/1_biomarker_association_workflow.ipynb
   

Documentation

For detailed documentation, visit our [Documentation Site](https://llamasearch.ai or check the docs/ directory.

• User Guide: Instructions for installation, configuration, and usage
• API Reference: Detailed documentation of modules, classes, and functions
• Examples: Jupyter notebooks demonstrating LlamaCancer workflows
• Tutorials: Step-by-step tutorials for common tasks

Example Results

Kaplan-Meier plot showing event-free survival stratified by B-cell gene expression signature.

Forest plot showing hazard ratios for multiple biomarkers.

Citation

If you use LlamaCancer in your research, please cite:

LlamaGroup. (2023). LlamaCancer: A framework for biomarker association analysis in oncology clinical trials.
GitHub repository: https://github.com/llamagroup/llamacancer

License

LlamaCancer is released under the MIT License. See the LICENSE file for details.