台灣種豬場場內檢定杜洛克豬隻生長及屠體性狀之遺傳評估

Abstract

本研究旨在評估台灣種豬場內檢定杜洛克豬隻的生長與屠體性狀之遺傳表現。本研究分析了2014至2024年間來自11個特定種豬場的14,948頭杜洛克豬隻性能資料，主要分析的生長與屠體性狀包括達結束檢定日齡（tage）、平均隻日增重（ADG）、平均背脂厚度（BF）、腰眼面積（LMA）、瘦肉率（Lean%），以及多個超音波性狀如第4肋背脂厚度（P1）、最後肋背脂厚度（P2）、最後腰椎背脂厚度（P3）、第10肋背脂厚度（P10）與實際腰眼深度（tLD），並採用線性混合模型與最佳線性無偏估測方法估計各性狀的遺傳參數與育種價。研究結果顯示，tage、ADG、BF、LMA、Lean%、P1、P2、P3、P10與tLD的遺傳率估計值分別為0.315、0.322、0.393、0.199、0.208、0.262、0.333、0.398、0.334與0.212。台灣杜洛克豬群的生長與屠體性狀多呈現中等遺傳率，與國際文獻報導相似，然而LMA、Lean%與tLD的遺傳率估計值偏低，可能因採用A模式超音波儀器進行間接測量，導致非遺傳變異增加所致。在改進趨勢方面，分析結果顯示在2014至2024年間，年平均修正表型數值只有BF、P2、P3與P10（P < 0.001）呈現顯著的改進趨勢，其餘性狀受環境干擾較大，變動趨勢未達統計顯著；除了LMA以外，其餘性狀的年平均育種價估測值則呈現顯著的改進趨勢（tage、ADG、BF、P1、P2、P3與P10皆為P < 0.001，tLD為P < 0.05），顯示在排除環境因素之干擾後，育種價的選拔效果得以實現，改進趨勢也符合選育方向。本研究建議現場選拔應以育種價估測值為依據，而非受多種非遺傳因素影響的表型數值，以提升遺傳改進效率；同時，透過改善飼養管理與降低環境變異，可進一步提升台灣種豬之選拔效率與整體生產效益。未來種豬場場內檢定改採B模式超音波儀器進行屠體性狀測定，亦可望提升LMA、Lean%與tLD等性狀之遺傳率估計值，增進選拔準確性。

This study aimed to evaluate the genetic performance of growth and carcass traits for duroc pigs in the Taiwan Swine On-Farm Test. A total of 14,948 performance records were collected from 11 herds between 2014 and 2024. The analyzed traits included age at test (tage), average daily gain (ADG), backfat thickness (BF), loin muscle area (LMA), lean percentage (Lean%), and several ultrasound traits: backfat thickness at the 4th rib (P1), backfat thickness at the last rib (P2), backfat thickness at the last lumbar vertebra (P3), backfat thickness at the 10th rib (P10), and tested loin depth (tLD). Genetic parameters and estimated breeding values (EBVs) were obtained using linear mixed models and the best linear unbiased prediction (BLUP) method. Heritability estimates for tage, ADG, BF, LMA, Lean%, P1, P2, P3, P10, and tLD were 0.315, 0.322, 0.393, 0.199, 0.208, 0.262, 0.333, 0.398, 0.334, and 0.212, respectively. The results indicated moderate heritability for most growth and carcass traits, consistent with values reported in international literature. However, LMA, Lean% and tLD showed relatively low heritability, which may be attributed to increased non-genetic variation associated with the use of A-mode ultrasound devices for indirect measurement. From 2014 to 2024, annual changes in adjusted phenotypic means were statistically significant only for BF, P2, P3, and P10 (P < 0.001), whereas other traits were more affected by environmental variation and did not reach statistical significance. In contrast, annual mean EBVs showed significant improvement trends for all traits except LMA (tage, ADG, BF, P1, P2, P3, and P10: P < 0.001; tLD: P < 0.05). These results confirm the effectiveness of selection based on EBVs when environmental effects are accounted for, and show that the improvement trends are consistent with the established breeding objectives. It is recommended that selection decisions be based on EBVs rather than phenotypic values, which are more susceptible to substantial non-genetic variation. Additionally, improving herd management and reducing environmental variability are essential to further enhance selection efficiency and overall productivity in the swine industry in Taiwan. The future use of B-mode ultrasound for carcass trait measurement in the Taiwan Swine On-Farm Test is expected to increase the heritability estimates for traits such as LMA, Lean%, and tLD, thereby improving selection precision.