Skip to content

sky-farmer/meBLUP

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 

Repository files navigation

README

Model equation BLUP (meBLUP) is a new method to estimate genetic breeding values (GEBVs) combining relationships form pedigree and genotype information. In comparison to traditional methods, it is not necessary to set up any relationship matrix or its inverse. meBLUp was developed by E. Groeneveld and A. Neumaier. For further information check the publications listed below.

The current repository includes the scripts developed and used in the Master's thesis of Himmelbauer, 2022 [2].

Publications

[1] E. Groeneveld and A. Neumaier. "BLUP without (inverse) relationsship matrix". In: World Congress on Genetics Applied to Livestock Production. Electronic Poster Session-Theory to Application 3. Auckland, New Zealand, 2018, p.21. URL: http://www.wcgalp.org/system/files/proceedings/2018/blup-without-inverse-relationship-matrix.pdf

[2] J. Himmelbauer. "Parameter estimation in model equations for animal breeding". Master's thesis, University of Vienna, 2022. URL:

[3] A. Neumaier and E. Groeneveld. "Model equations for prediction with pedigree". Manuscript in preparation. 2022.

Description

meCODE

The folder meCODE contains all R-scripts to run meBLUP as it is described in the the Thesis Himmelbauer, 2022 [2]. The file

meSTART.r

can be used to sstart a simple evaluation, estimating GEBVs, SNP-effects and varaince parameters (e.g. $h^2$) simultaneously.

The file

meSTART_EBV.r

contains the whole appraoch developed and tested in [3]. This appraoch is based on several steps and it can be used to estimate GEBVs, SNP-effects and some varaince parameters (e.g. polygenetic effect), if the heritability ($h^2$) or an estimator for $h^2$ is available for the respective dataset.

In these two start-Scripts several functions are used, wich are defined in the scripts provided in the folder meMML. For detailed information please read [2] and the description in the header of the respective R-script.

To run the start scripts, it is necessary to download the whole folder meCODE and save it localy. The start scripts have to be executed in the folder meCode. Otherwise the Variable PATH_meBLUP in the scripts has to be replaced with the full path to the folder meMML.

Toy-Example

In the folder meCODE/toy_data a very small toy dataset is provided. The two scripts can be tested on this toy dataset with the following code lines, assuming a $h^2$ of 0.3 and using all 4 SNPs. For all other parameters the default values are used. To run a single meBLUP evaluations this script can be used:

R CMD BATCH '--args "toy_data/toy" "0.3" "all" '  meSTART.r

To applay the developed approach to estimate EBVs the following script is used:

R CMD BATCH '--args "toy_data/toy" "0.3" "all" '  meSTART_EBV.r

meSIM

The folder meSIM contains all R-scripts to simulate a dataset, taht can be used for meBLUP. The simulation was used in [2] to get an independent dataset for test purpose. The script start.sh can be used to start the simulation:

$ sh start.sh

All parameters for the simulation can be changed in scripts/Sim.r. The parameters to change are explained in this file.

NOTE: The calculation of EBVs is done with the mix99 program suite (MiX99 Development Team. "Solving Large Mixed Model Equations". Jokionene, Finnland, 2019. URL: https://www.luke.fi/en/services/mix99-solving-large-mixed-model-equations). If these programs are not available, the scripts to estimate (G)EBVs for the simulation have to be adapted.

License

The code is licensed with GNU GENERAL PUBLIC LICENSE Version 2 (https://www.gnu.org/licenses/old-licenses/gpl-2.0.en.html). So everybody can freely use (even commercially) and modify the code, under some conditions:

  • give appropriate credit (associated Master's thesis and associated publications),
  • state if and what changes were made and
  • publish your work under the same public license as the original.