Improved metodology for analyzing survival data in fish breeding programs

Selection for increased disease resistance based on survival data obtained in commercial farm environments in inefficient due to low genetic variation for such survival traits. Higher genetic variation has been observed for specific disease resistance obs erved as survival in controlled challenge tests. However, the current statistical methodology used to obtain estimates of genetic variances and prediction of family breeding values for such challenge test survival data (a normal univariate mixed linear mo del) does not utilize all information in the data. The objective of this project is to evaluate available statistical methods developed for genetic evaluation of livestock (i.e. threshold and survival models) that will utilize more of the information in the observed survival data to better describe the genetic variation in disease resistance and thus improve the selection for increased disease resistance in fish breeding programs. Existing data from Atlantic salmon, Pacific prawn and Nile tilapia will be used when evaluating available statistical models. However, the mothodology will also be suitable to analyze survival data for traits in other aquaculture species and for other tra its that may be recorded in challenge tests (e.g. cod-water and salinity tolerance).