Genomic prediction using epigenetic profiles

Genomic prediction is a subfield of quantitative genetics that focusses on breeding value estimation, phenotype prediction, and assessment of genetic propensities to disease based on genomic information. Currently, it relies mostly on SNP-chips, mixed model methodology and recurrent model training of genotypic effects within animal or plant populations. At the same time, as biotechnology companies provide new methods and products to analyse cells and molecules, new data from other “omics” (e.g., epigenomics, transcriptomics) are becoming increasingly available for analysis, in greater volume and speed, and at reduced costs. The field of epigenetics has gained great popularity and academic interest, as it promises to increase our understanding of biological phenomena by extending the concept of heredity and genetic variation beyond “hard” base pair changes to include “soft” changes due to, for instance, DNA methylation. However, a lack of quantification of epigenetic effects on phenotypic variation is a setback for this accomplishment. It is necessary to offer new quantitative genetic models for incorporating epigenetic profiles in genomic prediction, as well as to test existing ones. We aim to quantify the importance of the contribution of epigenetics to quantitative trait variation by estimating variance components, which inform us how much phenotypic variation may be attributed to different variation sources, such as epigenetic profiles, genotypes, and environments. The project aims to develop and validate new models for genomic prediction integrating genetic and epigenetic profiles, especially DNA methylation profiles. The suggestion of new models will occur together with the development of scripts to implement the models on real and simulated data. Models will initially be tested with simulated and publicly available data. Finally, we will test the models with sperm motility traits provided by Topigs Norsvin Research Centre. We expect that the incorporation of epigenetic profiles will increase the accuracy of genomic prediction. This research project aims to test epigenomic models in a unique real data set, as well as to build upon those models to enhance the accommodation of methylation effects into prediction. The phenotypes, genotypes, and methylation profiles obtained in the GEroNIMO project will provide a special opportunity to test the different epigenomic prediction models that have been proposed. These aims will be addressed by pursuing the following two main objectives: first, develop and evaluate new models for genomic prediction integrating SNPs and epigenetic profiles, especially methylation profiles; second, estimate variance components of phenotypic observations attributed to epigenetics.