CONCLUSIONS

1. MATERIALS & METHODS (con’t.) • Analyses of CHIgroups Analyses were conducted by REML within each of the 8 CHI groups using the model: DO = HY + Year-State-Month + Lact + PE + A + e, Where HY is herd-year, year and month of calving, Lact is parity number, PE is permanent environment effect, A is additive genetic effect, and e is error effect. The genetic correlation between DO in synch and non-synch herds was estimated using the program MTDFREML. Herds were randomly selected within CHIgroups 4 through 8 for inclusion in statistical analysis. Over half the herds had 3 or fewer years of available data. Table 1 gives the criteria, number of herds, cows, and lactations for the 8 CHI groups. • Table 1. Data included in statistical analyses. • RESULTS (con’t.) The genetic correlations between DO for cows in CHI group 8 (non-synch) and DO for cows in the other 7 groups were near 1.0, indicating that DO can be considered the same trait regardless of synch. A further question is whether the differences in variances in Table 2 should be considered in genetic evaluations of DO/Pregnancy Rate. Preliminary results indicated only a slight improvement in sire PTA when DO records were adjusted for these differences. • INTRODUCTION • Use of synchronized breeding regimes is steadily increasing. • While variations in protocols exist, many provide for timed AI rather than observation of estrus behavior. • Fertility traits may be different genetically compared to traditional insemination based on observed estrus because expression of estrus behavior is no longer involved. • To examine this question requires a method to identify herds practicing synchronization (synch). • OBJECTIVES • Develop a method to identify herds which practice (a) synchronization/TAI (b) insemination based on observed estrus. • Estimate genetic and error variances for days open (DO) for alternative breeding practices. • Estimate the genetic correlation between DO with and without synchronization • MATERIALS & METHODS • Holstein cows calving between 1995 to 2005, required to have a 1st lactation and included AI 1st services only. • Assumed that herds practicing synch would deviate from an expected equal distribution of services among weekdays Sunday through Saturday. • Synch vs. Non-synch Herd Identification • Developed a chi2 based statistic (CHI) whose magnitude for inseminations in a herd-year was expected to increase with increased likelihood of the herd practicing synch. • The statistic is: • where Ni = number of inseminations in the herd-year performed on the ith weekday (i=1,7), and HS is herd size. Herd-years were grouped into 8 categories based on CHI and defined in Table 1. • CONCLUSIONS • Developed a method to identify herds synchronizing based on frequency of first inseminations by day of week. • Estimated genetic, PE, and error variances for DO within 8 CHI groups based on likelihood of practicing synchronization. • Genetic variance in DO was greater in herds following traditional estrus detection than in herds practicing synchronization. • Heritability of DO was higher (3.6%) in herds not synchronizing than in herds synchronizing (2.6%). • Repeatability tended to be higher in herds not synchronizing than in herds synchronizing. • Genetic correlation between DO in herds synchronizing and DO in herds not synchronizing was near 1.0. • Future research will future explore how heterogeneous variance adjustment affects genetic evaluations for DO/DPR. RESULTS Table 2. Single-trait variance estimates by CHI groups. • Synchronization (groups 1-6) lowered genetic variance and heritability compared to non-synchronized DO (groups 7 & 8). Days to first breeding (DFB) has a higher h2 than conception rate. Thus, synchronization is likely removing some genetic variance (in DO) related to DFB. • Error variance also tended to be slightly higher where a greater degree of synchronized breeding (groups 1 through 4) was occurring. • Synchronization generally lowered repeatability as well, although trend was less clear than for h2 7 CHI = 100 * { [Ni – E(Ni)]2/E(Ni)}/HS, i=1