Nitrogen efficiency, i.e. the proportion of nitrogen uptake in the body to the total nitrogen intake is an ecologically and economically important phenotype for the genetic selection of pigs. Preliminary analyses suggest a heritability of 32% and 16% for nitrogen efficiency in the empty body and carcass (1), respectively. However, a reliable estimation of genetic parameters or genome-wide association studies requires a large number of phenotyped individuals. Traditionally, the lean meat content of the carcass or the empty body is determined via dissections, which suffer from operator bias, or chemical analysis, which is labour-intensive and expensive. Thus, methods with a potential for high throughput are needed. These methods will also help develop precision feeding strategies. Here, we explore whether dual-energy X-ray absorptiometry (DXA) is a suitable alternative in terms of accuracy and precision in a Swiss Large White pig population. DXA scans provide a measure for body composition such as lean meat content (2), from which the nitrogen content can be derived. However, the values obtained by DXA are systematically biased because they are an indirect measure and therefore require calibration. Hence, prediction equations are needed to estimate the true nitrogen content (as obtained by chemical analysis as the golden standard). Methods: We obtained measurements of both nitrogen content by chemical analysis as previously described (3) and lean meat content with DXA (GE Lunar iDXA with pencil beam) of both live animals and carcasses of the same 68 entire males (6 pigs with live weight of 20 kg, 18 with 60 kg and 44 with 100 kg). . We derived linear equations to predict nitrogen content from DXA lean tissue measurements and we determined R2 and RMSE as measures of precision and accuracy. Results: We found that chemically determined nitrogen content can be predicted by DXA measurements with high accuracy and precision. The linear equation to estimate nitrogen content of the live pig is and for the carcass . Both equations yield high precision of estimates (R2=0.982 and R2=0.983 for live pig and carcass, respectively) and values could be predicted with little error as indicated by RMSE=4.7% and RMSE=4.4% of the mean of the chemically determined nitrogen content). Conclusions: Due to the high accuracy and precision of predicted nitrogen contents of body and carcass, we conclude that this method provides a practicable and non-invasive option for determining nitrogen content in empty body as well as the carcass in a high throughput fashion that is needed for genetic studies. The applicability of prediction equations is expected for females and castrates, but awaits confirmation. Since potential breeding animals do not have to be sacrificed, DXA could aid performance testing and the selection of parent individuals for breeding because breeding values can be estimated directly for dams and sires and not from relatives, which will increase accuracy and accelerate genetic gain. Non-destructive methods to determine body composition precisely will aid the development of precision feeding. Bibliographical references 1. KASPER C., RUIZ-ASCACIBAR I., STOLL P., BEE G. (2019): Agrarforschung Schweiz 10: 164-171 2. MITTCHELL A.D., SCHOLZ A.M., CONWAY J.M. (1998): Appl. Radiat.Isot. 49(5/6): 521-523 3. RUIZ-ASCACIBAR I., STOLL P., KREUZER M., BOILLAT V., SPRING P., BEE G. (2017): animal 11(3): 394-404
