Prediction of Son's Modified Contemporary Comparison from Pedigree Information

Abstract

Son's modified contemporary comparison Predicted Differences and modified contemporary deviations based on daughters' first lactations only and on all lactations were predicted for milk and fat yield and fat percent from pedigree information. Pedigree records included Predicted Differences and modified contemporary deviations for male relatives and modified contemporary deviations and Cow Index on first available records for female relatives. For a son's dam, however, first two and first three available records also were used. Models were chosen by a stepwise procedure with stopping when there was a change in the multiple correlation coefficient squared of less than 1%. Minimum repeatabilities of 40 and 70% for sons were required also to form two data sets.Models for milk and fat included sire's and maternal grandsire's Predicted Differences and dam's Cow Index. Models for fat also included the modified contemporary deviation for the maternal grandsire. For fat percent, sire's Predicted Difference or modified contemporary deviation and dam's Cow Index were important. Multiple correlation coefficients ranged from .32 for fat percent and .31 for milk to .24 for fat. Two-variable models suggested that for milk and fat, maternal grandsire was more important than dam whereas dam was more important for fat percent. The dam's Cow Index on first available record was more important for milk and fat, but dam's Cow Index from first three available records was most important for fat percent. Results were similar for predicting son's modified contemporary deviations. Closer prediction of son's Predicted Differences to real Predicted Differences for yield of milk and fat could be accomplished by placing slightly more emphasis on the maternal grandsire, and for fat percent, by putting more emphasis on dam's later records. ABSTRACTSon's modified contemporary comparison Predicted Differences and modified contemporary deviations based on daughters' first lactations only and on all lactations were predicted for milk and fat yield and fat percent from pedigree information. Pedigree records included Predicted Differences and modified contemporary deviations for male relatives and modified contemporary deviations and Cow Index on first available records for female relatives. For a son's dam, however, first two and first three available records also were used. Models were chosen by a stepwise procedure with stopping when there was a change in the multiple correlation coefficient squared of less than 1%. Minimum repeatabilities of 40 and 70% for sons were required also to form two data sets.

Prediction of Progeny Genetic Evaluations from Simultaneous Genetic Evaluations of the Dam, Sire, and Maternal Grandsire with an Animal Model

“…Westell and Van Vleck (17,18) summarized studies using Northeast AI sire comparisons (NEAISC) and intraherd cow evaluations (1,7) in the Northeast (NE ETA). Similar studies have been done with USDA evaluations (2,3,4,5,9,10,12).…”

“…Many reports have suggested that dam information does not predict son or daughter genetic value as well as theoretically possible (2,4,5,6,9,10,12,17,18). Evaluations for dams that include other than first lactation records have produced regressions for offspring on dams that are much smaller than genetic theory predicts.…”

Theoretical Weights for Regression of a Son's Genetic Evaluation on His Sire's and His Dam's Genetic Evaluations

“…The use of the maternal grandsire model (7) complicates the theoretical regression but will not be considered here because sire evaluation in pedigree analyses previously have not considered the maternal grandsire model (1,2,4,5,6,8).…”

Repeatability and bias of estimated breeding values for dairy bulls and bull dams calculated from animal model evaluations

“…However, even if preferential treatment could cause the heterogeneity, the use of within-herd variance or within-herd coefficient of variation has been found to be insensitive in detecting herds with preferential treatment (Graham, Smith and Gibson, 1991). Using only the first records of cows for bull dam selection, the overestimation could be reduced (Rothschild, Douglass and Powell, 1981;Pedersen, 1991). On the other hand, the accuracy of bull dam evaluation would then be reduced.…”