Comparing predictions of mean performance and environmental sensitivity of recombinant inbred lines based upon F3 and triple test cross families

• Published in: Heredity 1980-12, Vol.45 (3), p.305-312
• Main Authors: Jinks, J L, Pooni, H S
• Format: Article
• Language: English
• Subjects: Biomedical and Life Sciences; Biomedicine; Cytogenetics; Ecology; Evolutionary Biology; Human Genetics; original-article; Plant Genetics and Genomics
• ISSN: 0018-067X; 1365-2540
• DOI: 10.1038/hdy.1980.73

Summary We have previously shown that statistics estimated from the F 1 , F 2 , B 1 and B 2 families of a cross between two inbred lines or from a triple test cross on the F 2 of such a cross can be successfully used to predict the properties of the pure breeding lines extractable from the cross by single seed descent. In this paper we extend the method to properties such as mean performance and environmental sensitivity that can be measured only by replicating each family or generation over two or more contrasting environments. The optimal design for making predictions for such characters is the families of the F 2 triple test cross replicated over environments. The simplest design, which assumes that dominance and other non-additive sources of variation are not large, is F 3 families replicated over environments. These two designs have, therefore, been compared using the cross between varieties 2 and 12 of Nicotiana rustica , two sowing dates a month apart as the environmental treatment, and mean performance and environmental sensitivity in respect of flowering time and final height as the properties to be predicted. The predicted properties of the pure breeding lines that would be derived from this cross, both singly and jointly, are compared with those of a random sample of its F 7 families produced by single seed descent. There is no difference between the predictions based on the triple test cross and F 3 families, and the observed properties in respect of final height do not differ from either. For flowering time the agreement is not so good, the most likely cause being the unreliability of estimates based upon only two environments. Should the number of environments be increased to overcome this possibility or to cover a more realistic set of environmental variables of importance to the growth of the crop it can only improve the attractiveness of the simpler F 3 design over the triple test cross.