XSim version 2: simulation of modern breeding programs

Area of Science:

• Agricultural Science
• Computational Biology
• Genetics

Background:

• Breeding programs require efficient tools for design, evaluation, and optimization.
• Modern agriculture benefits from simulators integrating big data and advanced statistical models.
• XSim Version 1 validated genomic selection but lacked comprehensive features for modern breeding.

Purpose of the Study:

• To present XSim Version 2, an extensively redesigned, open-source simulation tool for modern breeding programs.
• To introduce enhanced features for integrating multiple statistical models and complex mating schemes.
• To demonstrate the tool's flexibility in simulating crossbreeding and incorporating biotechnology.

Main Methods:

• Redesigned XSim tool with a modular approach for simulating successive generations.
• Incorporation of multiple statistical models: GBLUP, Bayesian alphabets, and neural networks.
• Implementation in Julia for a balance of readability and computational speed.
• Overloaded functions and operators for an improved user interface and data management.

Main Results:

• XSim Version 2 seamlessly integrates diverse statistical models for genetic evaluations.
• The tool effortlessly simulates complex mating schemes and successive generations.
• Case studies demonstrate effective simulation of crossbreeding in animal and plant populations.
• Integration of modern biotechnologies like double haploids and embryo transfer is supported.

Conclusions:

• XSim Version 2 is a flexible, extensible, and user-friendly simulation tool for modern breeding programs.
• The simulator bridges the gap between theory and experiment in breeding research.
• Its design in Julia offers maintainability and extensibility for the scientific community.