Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 6/22/2003
Publication Date: N/A
Citation: Interpretive Summary:
Technical Abstract: National genetic evaluation for female fertility was implemented in February 2003. The evaluations are reported as daughter pregnancy rate. Pregnancy rate is calculated from days open (DO) as (233 - DO)/4. Currently, records must have a minimum of 250 days in milk (DIM) to be included for genetic evaluation. Furthermore, DO is set to 250 for records that go beyond that upper limit. This research examined the possibility of using DO records prior to 250 DIM by predicting unknown records. The prediction model included the fixed effects of lactation and calving ease and linear regressions on age at calving, average of first three test-day milk yields, previous DO, previous number of services, and days to first breeding. Quadratic effects of age and milk yield were also included. To assess the utility of the predictions, 11 DO groups were formed by defining the first group as 70 days or less and subsequent groups in 20 day increments. The final group was defined as >= 250-d. Each record was included in each group. Within group, y-hat was defined as actual DO if actual DO was <= the upper limit for that group or projected DO otherwise. Bias, standard deviation of prediction errors, and phenotypic correlations between y-hat and actual DO were calculated for each group. Genetic correlations were estimated for the groups with 90, 130, and 170 day upper limits. Bias ranged from -30 d (70-d group) to 0 d and was close to 0 d starting with the 110-d group. Standard deviation of prediction errors ranged from 49 d to 14 d. Phenotypic correlations increased from .41 (70-d) to .98 (250-d). Estimates of genetic correlations were 1 in all 3 groups examined. These results suggest that DO records can be utilized prior to the current 250-d requirement. Projected records require a weight less than one in genetic evaluation. Weights can be determined from correlations between actual and predicted records. Pregnancy confirmation code, now being collected, will also contribute to determination of weights.