Fertility, mortality and survival in Holsteins
Production of Holsteins increases but fertility, survival
and mortality deteriorate. The deterioration seems to be especially strong under
stressful conditions, including the heat stress. We are looking at G x E, particularly
under heat stress, and at nonlinear relationship between the major traits. One
of the current topics is evaluations with the availability of SNP data. Read
below.
Joint analysis of phenotypic,
pedigree and genomic data
Currently, genomic predictions using the data from the SNP panels are derived via a multiple step approach. These include a) regular genetic evaluation, b) calculation of unregressed proofs or daughter yield deviations, c) running predictions for animals with the genomic information, and d) combining predictions from regular and genomic evaluations. We are working on a methodology where all of these steps can be combined into one genetic evaluation. In particular, existing animal-model program may be modified to accept an “additive effects with genomic information”, and the evaluations could be run with the same model as currently.
Multibreed evaluation in beef cattle
We evaluate upgrading populations of beef cattle for a number of traits, including categorical while accounting for breed differences and heterosis. Models of interest include those for analyses at any age from birth to mature. UGA provides software for genetic evaluation to the U.S. beef industry.
Efficient yet simple animal-breeding programming in Fortran 90
Use of object-oriented and matrix operations in Fortran 90 can lead to
programs that are almost as simple as in a matrix language but as efficient as
in Fortran 77. Read a paper titled: Complex models, larger
data, simpler computing? This project has resulted in a large number of
application programs.
Evaluation for heat stress in dairy cattle
Whenever there is a heat stress, for example in
Economically efficient pig breeding
Although modern pigs grow faster and have more piglets, they also have higher
piglet mortality, lower sow survival, and higher susceptibility to diseases.
Some claim that no real genetic progress in pigs at the commercial level over
the past 10-20 years. In this project, we are looking how to select a
commercial animal that would have a good balance of fitness and production. One
complication is that important traits are censored and/or categorical. We also look at competition models.
Use of random regression models for a number of problems
· continuous growth,
· effect of aging on conformation scores,
· changing definition of a trait with time,
· changing genetic correlations with time.
We found that computations with random regression models can be greatly simplified if linear splines are used; one benefit is that variance components are on the same scale as in multiple trait models.
These and other projects can benefit from your collaboration. If you are looking for a graduate school or a place for a sabbatical, please consider the University of Georgia.
last updated Dec 22, 2008