Federated GWAS screening
Coordinate federated quality control, KING-based relatedness screening, and association screening across participating clients so that each site can shrink the set of loci that still need follow-up testing.
Start from the task in front of you
Core documentation paths are split by workflow so setup, examples, and reference material stay easy to scan.
Prerequisites
Set up Python, PLINK, Flower, and the required runtime tools.
Local Simulation
Run the default tiny correctness experiment from data generation to evaluation.
Examples
Browse tiny correctness, performance small, and 1000 Genomes workflows.
API Reference
Look up Flower run config, YAML schema, output files, and Python entry points.
Built for research operations
The documentation follows the repository structure so researchers can move from setup to pipeline internals without switching mental models.
Privacy-preserving coordination
Keep genotype-level computation local while using encryption, shuffling, anonymization, and lightweight secret-sharing. The server relays selected protocol messages without decrypting them.
Experiment-ready layout
Run repeatable Flower simulation and local-deployment experiments with scenario-specific configs, logs, and result directories.