An analytical approach to selecting anatomical sites for predicting in vivo body composition in pigs and goats using computed tomography derived tissue thickness measurements

This analytical approach aimed to determine anatomical sites from which in vivo whole-body composition of pigs and goats can be predicted, based on computed tomography image analysis. This approach included three scripts that enables the automated extraction of pixel matrix and its conversion to excel format for the calculation of fat and lean thincknesses across each of the 154 pixel columns within a pixel matrix of a region of interest extracted from a tomogram. Finaly, this approach proposed a statistical analysis using R to evaluate the prediction of whole-body fat and lean volumes (excluding viscera) in pigs and goats, based on fat and lean thicknesses measured by computed tomography and liveweight. Simple linear regressions were computed with the lm function. These scripts are associated to the method article entitled “Selection of anatomical sites for predicting in vivo body composition in pigs and goats using tissue thicknesses measured from computed tomography images”. The data used in this study were previously obtained in two animal experimentations (goats : Lerch et al, 2021, https://doi.org/10.1016/j.ymeth.2020.06.014 and in pigs : Quiniou et al, 2025 https://hal.science/hal-05050335v1).