Imputação de parentesco genético e predição das capacidades combinatórias em Serrasalmideos

Abstract

This work was conducted with the objective of ascribing the genetic parentage in Serrasalmidea with unknown parentage and predict general and specific combining abilities. We acquired 96 fingerlings from two commercial fish farms, with 12 originated from each of the following genetic groups: pacu, pirapitinga, tambaqui, tambacu, tambatinga, patinga, paqui and piraqui. The animals were randomly distributed into 16 water tanks (500 liters) in a water recirculation system (28oC), in which they were reared until 495 days of age. The fingerlings were weighed, submitted to morphometric analysis and processed. To confirm the identity of the animals, two nuclear and one mitochondrial markers were used. The combination ability predictions of the analyzed variables were obtained considering the genetic composition information of the animals given by the molecular analysis, using mixed models with a mixture of normal distributions in order to ascribe parentage of the animals considered advanced hybrids by the molecular analysis. The combination abilities were obtained using methodology proposed by Griffing (1956a) considering the mixed model, with the environmental effects estimated by means of EBLUE and the genetic effects considered random, obtaining the EBLUP of the general and specific combination effects. To verify the agreement between the ascribed incidence matrixes, and these with matrix provided by the producer, we performed the Pearson correlation using the Mantel test. The molecular analysis showed that the producer was correct in only 48% of the identity of the animals, finding 27 other advanced hybrids. There was no significant correlation (P>0.05) between the incidence matrix provided by the producer and the matrix ascribed, which caused divergence in the combination abilities between both methods. Between the general combination ability matrixes (GCA) ascribed, we verified high correlation (r>0.70) for different traits. Considering the methodology in which the missed model of normal distribution mixture was used for ascribing genetic composition, the tambaqui presented higher GCA and specific combination ability (SCA) for the majority of the analyzed variables, therefore being the most important genetic group. The GCA was shown to be more important than the SCA for al analyzed variables.