Related Experiment Video
Updated: Jun 1, 2026
05:53
Candidate Gene Testing in Clinical Cohort Studies with Multiplexed Genotyping and Mass Spectrometry
Published on: June 21, 2018
Pre-selection of markers for genomic selection.
Torben Schulz-Streeck1, Joseph O Ogutu, Hans-Peter Piepho
1Bioinformatics Unit, Institute of Crop Science, University of Hohenheim, Fruwirthstrasse 23, 70599 Stuttgart, Germany. piepho@uni-hohenheim.de.
BMC Proceedings
|June 1, 2011
Summary
Pre-selecting genetic markers improves the prediction of genomic breeding values (GEBVs). Excluding markers with inconsistent effects enhanced prediction accuracy, demonstrating the benefit of marker selection in genomic prediction models.
Area of Science:
- Animal breeding and genetics
- Quantitative genetics
- Genomic selection
Background:
- Accurate prediction of genomic breeding values (GEBVs) relies on numerous genetic markers.
- Predictive accuracy can be compromised by markers with null or inconsistent effects across populations.
Purpose of the Study:
- To evaluate the impact of marker pre-selection strategies on the accuracy of GEBV prediction.
- To compare the performance of ridge regression and spatial models for genomic prediction.
Main Methods:
- Three marker pre-selection approaches were investigated.
- Four Best Linear Unbiased Prediction (BLUP) methods, including ridge regression and spatial models, were employed.
- Model performance was assessed using 5-fold cross-validation.
Main Results:
- Ridge regression and spatial models exhibited comparable performance.
- Marker pre-selection significantly improved prediction accuracy, increasing the correlation between GEBVs and true breeding values from 0.607 to 0.625.
- An extended ridge regression model accounting for heterogeneous variances further boosted predictive accuracy to 0.648.
Conclusions:
- Pre-selecting genetic markers is a beneficial strategy for enhancing GEBV prediction accuracy.
- Excluding markers with inconsistent cross-specific effects positively impacts genomic prediction.
- Advanced modeling approaches, such as heterogeneous variance ridge regression, can further refine predictive performance.
