In vivo assessment of fat composition in Senegalese sole (Solea senegalensis) by real-time ultrasonography and image analysis of subcutaneous fat S.R. Silva a, , C.M. Guedes a , P. Rema a , A.C. Batista a,b , V. Rodrigues c , N. Loureiro a , J. Dias c a CECAV - Universidade de Trás-os-Montes e Alto Douro, Departamento de Zootecnia, Quinta dos Prados, 5000-801, Vila Real, Portugal b CAPES Foundation, Ministry of Education of Brazil, Brasília DF 70040-020, Brazil c SPAROS Lda, Área Empresarial de Marim, Lote C, 8700-221 Olhão, Portugal abstract article info Article history: Received 6 April 2015 Received in revised form 3 February 2016 Accepted 4 February 2016 Available online 11 February 2016 An ultrasound and image analysis non-invasive and non-destructive method for predicting skin fat, let muscle fat, and whole body fat depots in live Senegalese sole (Sole senegalensis) was evaluated. The experiment was conducted with Senegalese sole of commercial market size (n = 46, mean weight 278 ± 78 g). The shes under a mild anesthesia were placed in a rectangular metallic container and were scanned with a real time ultrasound (RTU) scanner equipped with a linear array transducer of 7.5 MHz. For all sh, four cross-sectional RTU images of the dorsal anterior let was obtained. The RTU images were analyzed using ImageJ software to determine the sh subcutaneous fat (SF) at each section. After this, sh were killed, and the right dorsal let of each sh was dissected in muscle and skin. These two components and the remaining of the sh were ground and homogenized for chemical analysis. The skin fat, let muscle fat and whole body chemical fat contents were obtained. The relationships between the SF obtained from image analysis and skin fat, let muscle fat and whole body fat were computed using simple regression and partial least squares regression (PLSR). The skin and let muscle had lower fat than the whole body (1.5 and 4.5 versus 6.6%, respectively). Filet muscle presents the lowest fat depot (1.5%). The SF obtained from RTU and image analysis shows a mean value between 2.8 and 3.5%. All fat depots were signicantly correlated with SF obtained by RTU and image analysis, with R 2 values ranging from 0.707 to 0.851 for skin fat, 0.725 to 0.802 for let muscle fat and 0.664 to 0.760 for whole body fat. The PLSR models explained accurately the skin fat (R 2 = 0.875 and RMSEP = 0.506), let muscle fat (R 2 = 0.800 and RSMEP = 0.117) and whole body fat (R 2 = 0.785 and RSMEP = 0.586). Our results suggest that RTU and image analysis is a practical and feasible non-invasive and non-destructive technique for in vivo fat depots prediction in Senegalese sole sh. Statement of relevance: In aquaculture, the knowledge of the body composition in live sh, particularly fat content, is important for feeding, reproduction and genetics programs. However, all those programs depend critically on quick, accurate, and above all, non-invasive methods to predict body composition in live sh. Traditionally sh body composition was determined by comparative slaughtering followed by chemical analysis. Comprehensive studies using image techniques such as real time ultrasonography (RTU) show the ability of this technique to predict sh body composition traits. This technique is non-invasive and non-destructive and therefore is very useful for use in vivo for sh body composition examinations. Our present work investigated the ability of RTU and image analysis to predict in vivo fat composition of Solea senegalensis. Through this work monitoring fat variations with time and the possibility of repeated measurements on the same individuals and on a high number of live sh throughout the life-cycle is possible. Therefore our present study has close relevance to aquaculture particularly to nutrition status of farmed shes. © 2016 Elsevier B.V. All rights reserved. Keywords: Senegalese sole Subcutaneous fat Ultrasound Image analysis 1. Introduction Sole (Solea solea and Solea senegalensis) have been considered as promising new atsh species for Mediterranean marine sh farming due to the need to develop protable markets (Dinis et al., 1999, Imsland et al., 2003). S. senegalensis is usually raised in extensive polyculture (in earth ponds) in the south of the Iberian Peninsula, where it can achieve higher growth rates than European seabass, being second only to gilthead seabream (Dinis et al., 1999). Its high price and market demand have triggered the interest in producing this species under more intensive conditions. In fact, Senegalese sole production increased from 110 to 500 tons from 2008 to 2011, especially in Portugal and Spain. The knowledge of the body composi- tion in live sh, particularly fat content, is important for feeding, repro- duction and genetic programs (Probert and Shannon, 2002; Romvári Aquaculture 456 (2016) 7682 Corresponding author. E-mail address: ssilva@utad.pt (S.R. Silva). http://dx.doi.org/10.1016/j.aquaculture.2016.02.009 0044-8486/© 2016 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Aquaculture journal homepage: www.elsevier.com/locate/aquaculture