MiX99 – Solving Large Mixed Model Equations

Luke’s MiX99 software suite is developed for high performance breeding value and variance component estimation in large-scale genetic and genomic evaluations. MiX99 supports a wide variety of prediction models and datasets. Cutting-edge research in prediction methodologies and computing algorithms, in close collaboration with research and breeding industry partners, ensures a continuous adding of topical modelling features.  

Among the largest applications are national genomic evaluations with a massive number of genotyped animals and complex models like multiple-trait single-step random regression test-day models. During the last decade, the use of MiX99 has spread in research, and worldwide it has become one of the leading software packages for national and international evaluations in cattle and other livestock species.

Luke grants three different MiX99 user licenses for research or genetic evaluations exclusively in animal breeding applications and warrants optimal tailoring of the MiX99 executables to the operating system (Linux, Windows, macOS) and hardware of the user.

• MiX99 Standard license: The MiX99 Standard license includes the most recent officially released version of MiX99. This includes data pre-processor program mix99i, the multi-threading solving program mix99s, reliability approximation program apax99, reliability calculation program exa99, as well as the multi-process MPI programs mix99p and apax99p.  
• MiX99 Core license: The MiX99 Core license includes the full capacity of the MiX99 data pre-processor program mix99i, the solving program mix99s, and the reliability approximation program apax99. The MiX99 Core license is available for emerging countries that are establishing genetic evaluation technology and evaluation centres.  
• MiX99 University license: The MiX99 University license includes the same programs as the MiX99 Core license but comes with a reduced users support and is exclusively granted for academic use only.  

For research and industry partners who are contributing to the further development of the MiX99 software, a license is made available that goes beyond the MiX99 Standard license and gives access to MiX99 development versions that include future features under beta testing.

For further information MiX99 team can be contacted at mix99@luke.fi. 

MiX99 users are welcome to join our discussion forum “MiX99 User Community” in Teams. Send a request to napoleon.vargas@luke.fi or timo.j.pitkanen@luke.fi  and you will be added to the team.   

MiX99 methodology development 

Strandén, I., & Jenko, J. (2024). A computationally feasible multi-trait single-step genomic prediction model with trait-specific marker weights. Genetics Selection Evolution, 56(1), 58. https://doi.org/10.1186/s12711-024-00926-2 

Strandén, I., Mäntysaari, E. A., Lidauer, M. H., Thompson, R., & Gao, H. (2024). A computationally efficient algorithm to leverage average information REML for (co) variance component estimation in the genomic era. Genetics Selection Evolution, 56(1), 73. https://doi.org/10.1186/s12711-024-00939-x 

Gao, H., Kudinov, A. A., Taskinen, M., Pitkänen, T. J., Lidauer, M. H., Mäntysaari, E. A., & Strandén, I. (2023). A computationally efficient method for approximating reliabilities in large-scale single-step genomic prediction. Genetics Selection Evolution, 55(1), 1. https://doi.org/10.1186/s12711-022-00774-y 

Strandén, I., Matilainen, K., Aamand, G. P., & Mäntysaari, E. A. (2017). Solving efficiently large single‐step genomic best linear unbiased prediction models. Journal of Animal Breeding and Genetics, 134(3), 264-274. https://doi.org//10.1111/jbg.12259 

Strandén, I., & Mäntysaari, E. A. (2014). Comparison of some equivalent equations to solve single-step GBLUP. In Proceedings of the 10th World Congress on genetics applied to Livestock production. Vancouver (p. 22). 

Matilainen, K., Mäntysaari, E. A., Lidauer, M. H., Strandén, I., & Thompson, R. (2012). Employing a Monte Carlo algorithm in expectation maximization restricted maximum likelihood estimation of the linear mixed model. Journal of Animal Breeding and Genetics, 129(6), 457-468. https://doi.org/10.1111/j.1439-0388.2012.01000.x 

Vuori, K., Strandén, I., Sevón-Aimonen, M. L., & Mäntysaari, E. A. (2006). Estimation of non-linear growth models by linearization: a simulation study using a Gompertz function. Genetics Selection Evolution, 38(4), 343-358. https://doi.org/10.1186/1297-9686-38-4-343 

Strandén, I., & Lidauer, M. (2001). Parallel computing applied to breeding value estimation in dairy cattle. Journal of dairy science, 84(1), 276-285. https://doi.org/10.3168/jds.S0022-0302(01)74477-3 

Lidauer, M., Strandén, I., Mäntysaari, E. A., Pösö, J., & Kettunen, A. (1999). Solving large test-day models by iteration on data and preconditioned conjugate gradient. Journal of dairy science, 82(12), 2788-2796. https://doi.org/10.3168/jds.S0022-0302(99)75536-0 

Strandén, I., & Lidauer, M. (1999). Solving large mixed linear models using preconditioned conjugate gradient iteration. Journal of Dairy Science, 82(12), 2779-2787. https://doi.org/10.3168/jds.S0022-0302(99)75535-9