PyMSQ 0.1.2

A Python package for fast Mendelian sampling (co)variance and haplotype-based similarity in genomic selection

PyMSQ: a Python package for fast Mendelian sampling (co)variance and haplotype-based similarity in genomic selection

PyMSQ is an open-source Python package that enables breeders, geneticists, and quantitative biologists to estimate Mendelian sampling–related metrics—including variance, covariance, and haplotype-based similarities—in both plant and animal species. For simplicity, PyMSQ consists of a single module, msq.

Key Features

• Within-Family Covariance
  Constructs population-specific covariance matrices that capture within-family linkage disequilibrium, reflecting recombination patterns and phased marker data.

• Mendelian Sampling (Co)Variance
  Estimates Mendelian sampling variance (MSV) for single or multiple traits, as well as covariances (MSCs), crucial for maintaining genetic diversity and controlling inbreeding.

• Similarity Matrices
  Computes haplotype-based similarity matrices between individuals (or zygotes), focusing on shared heterozygous segments that drive within-family genetic variation.

• Selection Criteria
  Offers functions to derive selection strategies (e.g., GEBVs, usefulness criteria, index-based approaches) that leverage MSV/MSC or similarity measures.

Installation

PyMSQ is available on PyPI and can be installed via:

python -m pip install PyMSQ

Basic Usage

Below is a minimal example illustrating how to import PyMSQ and call its core functions:

from PyMSQ import msq  # Imports the msq module

# Example: Loading an included dataset
data = msq.load_package_data()

# Deriving expected LD matrices
ld_matrices = msq.expldmat(data['chromosome_data'], data['group_data'])

# Estimating Mendelian sampling (co)variances
msv = msq.msvarcov(
    gmat      = data['genotype_data'],
    gmap      = data['chromosome_data'],
    meff      = data['marker_effect_data'],
    exp_ldmat = ld_matrices,
    group     = data['group_data']
)

# Constructing similarity matrices
similarity = msq.simmat(
    gmat      = data['genotype_data'],
    gmap      = data['chromosome_data'],
    meff      = data['marker_effect_data'],
    group     = data['group_data'],
    exp_ldmat = ld_matrices
)

Tutorial

A tutorial detailing each function’s parameters, usage examples, and best practices can be found here. This tutorial walks you through:

1. Loading your own data or the bundled Holstein-Friesian dataset,

2. Building LD matrices for each chromosome,

3. Estimating Mendelian sampling (co)variances,

4. Deriving haplotype-based similarity,

5. Applying selection strategies using advanced metrics.

Citation

If you use PyMSQ in academic work, please cite the following papers:

1. Musa, A. A., & Reinsch, N. (2026). PyMSQ: a Python package for fast Mendelian sampling (co)variance and haplotype-based similarity in genomic selection. BMC Bioinformatics, https://doi.org/10.1186/s12859-026-06392-5.

2. Musa, A. A., & Reinsch, N. (2025). A similarity matrix for hedging haplotype diversity among parents in genomic selection. Journal of Animal Breeding and Genetics, https://doi.org/10.1111/jbg.12930.

3. Zenodo (software version v0.1.2): https://doi.org/10.5281/zenodo.18643470

Funding

This study was supported by the Bundesanstalt für Landwirtschaft und Ernährung (BLE) under Grant 281B101516.

Getting Help

• Issues & Feature Requests If you encounter bugs, have feature requests, or need additional clarification, please open an issue on the PyMSQ GitHub repository.

• License PyMSQ is released under the MIT License, allowing both academic and commercial use.

Happy analyzing! We hope PyMSQ supports your work in breeding, helping you balance short-term genetic gains with the long-term preservation of essential haplotype diversity.