COMPORTAMIENTO DE DOS MODELOS NO LINEALES PARA ANALIZAR EL CRECIMIENTO EN RATAS DIABÉTICAS. IMPORTANCIA DEL CRITERIO BIOLÓGICO Perfomance of Two Non Linear Models to Analyze Growth in Diabetic Rats. Relevance of Biological Criteria

This study aimed to analyze the growth of eSS, and eSMT genotypes of rats with spontaneous diabetes. To achieve this, the ability to describe of von Bertalanffy and Logistic mathematical was evaluated, to describe models body weight and tail length data from weaning until 70 days after birth. The estimates of parameters were calculated using an iterative technique for nonlinear regression based on Marquardt algorithm. Different criteria were considered to compare the functions. The values of R were higher than 89%. The small variation in SCE values between models indicates that regardless of genotype, the two curves properly adjusted the rats’growth. As for the Akaike criterion, the variation between models for the adjustment of the data was very narrow, indicating an adequate articulation of the functions used. However, conspicuous differences were found regarding the biological interpretation of the parameters. Thus, the logistic function was unable to detect differences between the estimates of the parameters, despite the real differences noted between genotypes. Instead, the model of von Bertalanffy allowed to distinguish different patterns of growth, indicating that eSS and eSMT develop, at a lower speed, a longer skeleton on which biomass increases comparatively more, while achieves higher weight more slowly and lower length much faster.