Zigo: Sex Checking by Zigosity Distributions

A command-line tool for sex inference from genomic data. This tool uses a distilled polynomial equation to predict genetic sex from SNP zygosity distributions and can compare predictions with provided PED files.

Installation

We are working on publishing this package to PyPI. In the meantime, you can install the latest version directly from the repository using pip:

pip install git+https://github.com/AI-sandbox/zigo.git

Key Features

• Genetic Sex Prediction: Accurately predicts genetic sex from genomic data using a machine learning model
• PED File Integration: Optional comparison with provided sex information from PED files (requires "Individual ID" and "Gender" columns)
• Comprehensive Logging: Detailed logs of the analysis process

Usage

zigo -i INPUT_FILE -o OUTPUT_DIR [--ped PED_FILE]

Arguments

• -i, --input: Path to the genomic input file (.vcf, .vcf.gz, .bed, .pgen)
• -o, --output: Directory for saving results
• --ped: (Optional) Path to the PED file containing 'Individual ID' and 'Gender' columns

Input Format Requirements

• Genomic Data: Supports .bed, .pgen, and includes a high-performance C-based reader for .vcf and .vcf.gz.
• PED File: Tab-separated file with at least two columns:
  • 'Individual ID': Sample identifiers matching those in the genomic data
  • 'Gender': Sex information coded as 1 (male) or 2 (female)

Output Files

• results.sexcheck: Main results file with sex predictions, containing the following columns:

  • FID: Family ID (same as IID in this implementation)
  • IID: Individual ID
  • SID: Sample ID (same as IID in this implementation)
  • PEDSEX: Sex as provided in the PED file (1=male, 2=female, NA=not available)
  • SNPSEX: Predicted sex from SNP data (1=male, 2=female)
  • STATUS: Comparison result ("OK" if PEDSEX matches SNPSEX, "PROBLEM" if they differ)
  • F: Probability score for the predicted sex label

  Columns related to Y chromosome have been skipped.

• results.nosex: List of samples with no sex information

• sexcheck.log: Detailed log of the analysis process

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Cite

If you use Zigo in your research, please cite the source code.

📝 Note: Preprint will be published soon! Once published, we will update this section with the official citation.

In the meantime, please cite the repository as follows:

BibTeX:

@misc{zigo_2026,
  author = {Oscar Molina-Sedano, Daniel Mas Montserrat and Alexander Ioannidis},
  title = {Zigo: Sex Checking by Zigosity Distributions},
  year = {2026},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{[https://github.com/AI-sandbox/sex-check](https://github.com/AI-sandbox/sex-check)}}
}