Desenvolvimento de um modelo preditivo para avaliação de filé e carcaça de Pacama (Lophiosilurus alexandri) a partir de medidas morfométricas

Abstract

Morphometric measures has been a useful tool in Aquaculture for the development of predictive models of weights and yields of edible parts of fish. However, the analysis the weight of fillet and carcass and the body yields involves the slaughter of the animal. Therefore, these models have great importance for the fishing industry and in breeding programs, allowing a direct selection with a characteristic of interest. The pacman catfish (Lophiosilurus alexandri) is an endemic specie of the São Francisco River and it has gained space in Aquaculture as a production species. The aim of this study is to develop regression models associating weights and yields of fillet and carcass with different morphometric variables. The pacman catfish presented average yield of fillet (48.73%) and carcass (61.18%). These values are higher than those usually found for catfish species. For the construction of the predictive models, a variable weight allowed a construction of the models with the largest R2. An interaction between the morphometric variables, standard length and width at the insertion of fins, allowed the construction of well-adjusted models only after the transformation of the response variables. No correlation was observed between body yields and morphometric variables, and it was not possible to construct regression models for these variables.