Animal breeders have been very successful in improving the phenotypes (outward appearance) of animals, but knowledge about which genetic variants and genes exactly contribute to the observed phenotype is still anecdotal. Scientists from Animal Breeding and Genomics of Wageningen University & Research (WUR) and Topigs Norsvin have succeeded in identifying causal genes and variants by using a new machine learning tool called pCADD. Their findings were published in the Genomics journal.
Animal breeders have been very successful in improving their stock using a process called genomic selection. Genomic selection uses a set of predefined genetic markers to calculate a breeding value (the genetic potential of an animal). However, these genetic markers are not responsible for the change in phenotype, but the markers are genetically linked to the causal variants. Hence, despite the remarkable improvement of the animals, little is known about the molecular mechanisms (i.e. causal genes and variants) that are responsible for the remarkable changes observed in commercial livestock breeds. In this study, an approach was developed to identify causal variation and molecular pathways underlying important phenotypes in pigs.
To associate genomic regions with a phenotype of interest often genome wide association studies (GWAS) are performed. GWAS provide information on the location and significance of specific genomic regions associated with a phenotype. However, still many thousands of variants often underlie a specific GWAS result, while only one variant is usually causal. To pinpoint the causal variant, a new machine learning algorithm was developed that provides variant impact scores for any possible variant in the pig genome, allowing to rank the variants according to its likely importance. The efficacy of this approach was demonstrated by reporting known and novel causal variants, which provides a major step forward in underpinning the molecular mechanism underlying breeding in livestock.
Biology driven breeding (implications)
The newly developed approach allows disentangling the biology behind important phenotypes by accelerating the discovery of novel causal variants and molecular mechanisms affecting important phenotypes in pigs. This information will be used to improve breeding by adding the causal genes and variants to the marker panel used for genomic selection.