台灣荷蘭乳牛泌乳性狀之遺傳參數估值

Abstract

本研究之目的係利用台灣荷蘭乳牛群之現場資料，估計包括乳產量、乳成分與體細胞數分數等泌乳性狀之遺傳率與遺傳相關，此類重要經濟性狀之遺傳參數未來可應用愈台灣乳牛之遺傳評估作業。 依據台灣乳牛群性能改良計畫 (dairy herd improvement, DHI) 所建立之DHI資料庫，以SQL server之ODBC (open database connectivity ) 資料連線設定擷取牛隻初產日期介於1989年4月至2003年11月間計580場酪農戶、1,752頭公牛及已完成乳量性能檢定之後裔女兒牛共109,944頭，經由dBASE III PLUS程式與SAS統計分析軟體 (SAS 8.2, 1991) 進行資料編輯與整理，最後使用VCE5 (variance component estimate 5.0, 2003) 軟體進行前三泌乳期產乳量與生乳品質等性狀之遺傳參數估計，整理後各泌乳期的資料集分別包括性狀資料檔各27,917、18,759及9,258筆紀錄，固定效應變級數CYS (herd-year-season) 各361、358及339個與牛隻系譜資料檔包括5個世代，各45,651、30,095及19,257筆紀錄。 台灣荷蘭乳牛泌乳性狀遺傳參數之估算是以多性狀動物模式 (animal model) 之限制性最大似然法 (Restricted Maximum Likehood, REML) 進行。估計結果顯示，全期乳量 (305 days milk yield, MY)、全期脂量 (305 days fat yield, FY)、全期蛋白質量 (305 days protein yield, PY) 與體細胞數分數等性狀，於前三泌乳期之遺傳率估值分別為0.34 ± 0.02、0.37 ± 0.02、0.38 ± 0.02與 0.16 ± 0.01；0.22 ± 0.02、 0.21 ± 0.02、0.24 ± 0.02與0.09 ± 0.01；0.24 ± 0.03, 0.28 ± 0.03, 0.31 ± 0.03 與0.12 ± 0.03，各項乳產量性狀多為中至高度遺傳率；而體細胞數分數性狀則為低遺傳率，此結果與國外其它研究報告之估計相似。前三泌乳期之遺傳率估值以第一泌乳期最高，後續泌乳期則較低，此現象可能為環境變方隨泌乳期的增加而變大所致。在遺傳相關方面，前述之乳產量性狀間於前三泌乳期內均呈現高度之遺傳正相關，估值範圍介於0.82至 0.99之間，顯示選拔台灣荷蘭乳牛產乳量性狀對於增加乳脂量與蛋白質量具有正向選拔反應，惟體細胞數分數與乳產量性狀之間多為中度遺傳負相關 (–0.40至 0.13)，若進行單性狀選拔低體細胞數分數牛隻之策略可能有利於產乳量的提昇。

The objectives of this study were to estimate genetic parameters, including heritabilities and correlations, of milk yield, milk components and somatic cell score traits with field data of Holsteins in Taiwan, All these genetic parameter estimates of important economic traits would be expected to be used in the genetic evaluation of dairy cows in the future. Dairy production data were retrieved through SQL server using open database connectivity (ODBC) from the Dairy Herd Improvement (DHI) database in Taiwan, and original 109,944 records from Holstein cows with first calving from April 1989 to November 2003 were included which were from to 580 herds and 1,752 sires, respectively. These original retrieved data were edited with dBASE III PLUS and used to request summary statistics by SAS package firstly, and the genetic parameters of milk yield, somatic cell score and milk components for the first three lactations were estimated with VCE5 software. The data set included files of trait records (27,916, 18,758 and 9,257 cows for first, second and third lactation, respectively), county-year-season fixed factor (361, 358, and 339 levels for each lactation, respectively), and pedigree of 4 generations of ancestors. The genetic parameters of lactation were estimated with Restricted Maximum Likelihood using animal model. Model included fixed county-year-season effects, random animal and error. Estimated heritabilities for the first three lactations included 305d milk yield (MY), fat yield (FY), protein yield (PY) and log somatic cell score (LSCS) were 0.34 ± 0.02, 0.37 ± 0.02, 0.38 ± 0.02, and 0.16 ± 0.01 for first lactation, 0.22 ± 0.02, 0.21 ± 0.02, 0.24 ± 0.02, and 0.09 ± 0.01 for second lactation, and 0.24 ± 0.03, 0.28 ± 0.03, 0.31 ± 0.03 and 0.12 ± 0.03 for third parity respectively. Those heritability estimates were moderate for production traits and lower for log somatic cell score trait which was in agreement with the literature. Heritability estimates of the first lactation for all traits were higher than the later lactations been estimated, and the decreasing of heritability estimates for later lactations may be caused by increasing environmental variance. Genetic correlation estimates across lactations for the three milk yield traits were positive and ranged from 0.82 to 0.99, implying that selection for milk yield will appear genetic improvement for fat yield and protein yield traits. There were almost negative genetic correlations estimates between production and log somatic cell score traits (ranged from –0.40 to 0.13), however, LSCS was not show disadvantageous effect if only milk yield was selected.