Current projects
Use of semi-parametric multiple shrinkage methods in genome-assisted selection
Several new challenges in the analysis of genomic data have emerged with the availability of large panels of markers in livestock species. In QTL mapping studies, when all available markers are included in the analysis models become over-parameterized. The same problem arises in the context of genome assisted selection where precise estimates of each marker effect become problematic with oversaturated models. Model selection approaches as the ones proposed by Yi et al, (2003, and 2008) can represent a solution.
A different approach as outlined by Yi and Xu, (2008) it’s the use of shrinkage or regularization. One of the limitations of this approach resides in the utilization of priors with single mean and scale parameters.
It would be desirable to develop hierarchical priors to allow shrinkage of coefficients toward multiple prior means with unknown scale (Yi and Xu, 2008). Recently, MacLehose and Dunson (2008 ) proposed a multiple shrinkage semi-parametric method that allows shrinkage to multiple locations. In their method they utilize a Dirichlet process prior on mean and scale parameters which induce clustering around a small set of non-null means with different degree of shrinkage.
The objective of our stydy is to: 1) Investigate semi-parametric multiple shrinkage models in genomic evaluations of dairy bulls in comparison to methods currently employed. 2) Investigate the ability of these models to perform on a subset of SNPs. 3) expand semi-parametric multiple shrinkage methods to include a polygenic effects. 4) investigate the use of different prior specifications on the semi-parametric model.
The effect of population stratification in two-tier genomic-assisted selection
Under construction
Genetic variability of BVD vaccine response in cattle
This project is in collaboration with Dr. Michael Gonda at SDSU
It is oftentimes assumed that vaccine response is homogenous among animals, but the truth is that animals do not respond equally to administration of a vaccine. How much of the variation in the humoral response to vaccination is caused by genetic differences among animals?
We are initially focusing our efforts on the response to a commercially available bovine viral diarrhea (BVD) vaccine. BVD is a virus that can cause respiratory disease in cattle. If a cow is infected with BVD during pregnancy, that cow’s calf can become persistently infected with the virus and shed BVD rapidly to other animals in a herd.
Specifically, we are asking the following questions about the genetics of BVD vaccine response. 1) Can we identify breeds of cattle that are high or low responders to BVD vaccination? 2) What genes affect variation in BVD vaccine response? To answer this question, we will use a combination of whole-genome and candidate gene approaches. 3) What genes are up- and down-regulated following vaccination? To answer all of these questions, we will utilize techniques from several scientific disciplines, including quantitative genetics, molecular genetics, and functional genomics.
Investigation of parameters of milk release in the Italian Brown Swiss cattle population
This project is in collaboartion with Dr. Alessandro Bagnato at UNIMI (ITA) and Italian Brown Cattle Breeders' Association ANARB
Milkability traits have been indicated as one of the most important variable after milk yield and fat content, affecting farmer’s net profit. Data collected using portable flowmeters have been shown to improve milkability traits.
The use of a complete set of data, describing the overall milking pattern can prove to be an effective way to select animals that require lower management costs, improved udder health and consequently decreased veterinary costs associated with udder infection.
The current investigation involves both quantitative approaches to estiamte variance components for different parameters of milk release in the Brown Swiss population, and molecular approaches through genome-wide association aimed at identifying regions of the genome associated with these parameters
