Estimativa de parâmetros genéticos e predição dos valores genotípicos em heliconia spp. via modelos mistos

Abstract

Flower stem production were collected, in Germplasm Heliconia Collection of the University Federal Rural of Pernambuco (UFRPE), from December 2003 to May 2006, to estimate the genetic parameters and predicts genotypic values of Heliconia genotypes, using mixed models (REML/BLUP). The flower stem production were evaluated in: H. psittacorum L.f. x H. spathocircinata Aristeguieta ‘Golden Torch’ (GT) and ‘Golden Torch Adrian’ (GTA), Heliconia psittacorum L.f. ‘Suriname Sassy’ (SS), ‘Red Opal’ (RO), ‘Red Gold’ (RG), Heliconia x nickeriensis Maas e de Rooij (NCK), H. latispatha Bentham cv. Yellow Gyro (YG), H. rauliniana Barreiros (RB), H. rostrata Ruiz e Pavón (RRP), H. wagneriana Petersen (WP), H. bihai (L.) L. cv. Lobster Claw (LC). Statistical analysis was performed using the SELEGEN (55 model), to the genotype evaluation in completely randomize block with temporal stability and adaptability, in only one place and many harvest. In selecting genotypes for production and flower stalks per quarter, stood out GT, SS and RO genotypes. The quarters in which the genotypes were more productive were 6 (March to May 2005-18 MAP) and 9 (December to February 2006-27 MAP). Genotypes more stable were WP and WG. The genotypes better adapted were: GT, YG, GTA and SS. In a joint selection for flower stem production, stability and adaptability better performance were: GT, RO and GTA. The mean and variance components estimative, considering all genotypes in all quarters, were obtained. The variances have indicated strong influence of genotype x quarter in the phenotypic value. The ranking of genotypes was variable over the quarters due to genetic correlations considered average (43.52%). Regarding the genetic gains over general average in the combined analysis of the quarters, the genotypes GT, SS, RO and GTA were the best. Five quarters were required to evaluate the genetic potential, maintaining the heritability average of more than 70% with selective accuracy of 85%.This study identified the production of each genotype for each quarters, allowing better planning of marketing and chain supply. Furthermore, those results could guide other studies involving crosses to obtain new genotypes with superior characteristics, or even the other heliconia genotypes evaluation under different conditions, allowing the characterization of a large number of species.