A comparison of methods for ranking boars from different central test stations

Area of Science:

• Animal Breeding and Genetics
• Quantitative Genetics
• Swine Production

Background:

• Centralized boar testing stations are crucial for genetic improvement in swine populations.
• Accurate genetic evaluation methods are essential for selecting superior breeding stock.
• Inter-station connectedness is vital for robust genetic assessments across diverse environments.

Purpose of the Study:

• To compare the effectiveness of different genetic evaluation methods for ranking boars.
• To assess the impact of incorporating relative performance and reference sires on boar rankings.
• To determine the optimal statistical procedure for genetic merit evaluation in swine testing programs.

Main Methods:

• Analysis of performance data (average daily gain and backfat) from 3,999 boars across 24 test stations.
• Implementation of four ranking methods: individual records, contemporary deviations, and expected progeny differences (EPD) using a reduced animal model (RAM).
• Comparison of boar ranks using Spearman rank correlation to evaluate changes across evaluation complexities, including reference boars.

Main Results:

• Boar rankings changed significantly with increasing analytical complexity for both average daily gain and backfat.
• Rank correlations between individual records and contemporary deviations were moderate (0.47 for ADG, 0.20 for ABF).
• Rank correlations between contemporary deviations and RAM-based EPD were higher (0.53 for ADG, 0.41 for ABF), indicating improved accuracy with complex models.

Conclusions:

• More sophisticated genetic evaluation methods, such as EPD derived from RAM, provide more accurate rankings than simpler methods.
• The inclusion or exclusion of reference boars did not significantly alter the rankings derived from EPD.
• Adoption of best linear unbiased predictors is recommended for accurate boar comparisons in central testing facilities to maximize genetic gain.