Article

Trends and Implications of Genotype by Environment Interaction in South African Sugarcane Breeding

Journal of Crop Improvement 03/2012; 26(2):163-176. DOI: 10.1080/15427528.2011.622429

ABSTRACT Genotype by environment interaction (GxE) influences and complicates the selection of superior genotypes in trials by confounding the determination of true genetic values. In South Africa, variety trials are planted at several locations and harvested in the plant to third ratoon crops. The objective of this study was to determine the trends in components of GxE and their implications. The MIXED procedure of Statistical Analysis System (SAS) was used to estimate variance components. Genotype by location interaction was significant for the irrigated and coastal long-cycle programs, indicating the importance of identifying and characterizing sites. Genotype by crop-year interaction was larger and more significant for rain-fed than for irrigated cropping system, indicating the importance of ratooning ability in rain-fed regions. Genotype by location by crop-year interaction was significant (P < 0.01) for yield and sucrose content, highlighting the complexity associated with breeding sugarcane. The coastal long-cycle program was the most complex and generally characterized by large GxE. Separating the coastal hinterland and coastal average potential would be recommended to reduce GxE.

3 Followers
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Sugarcane (Saccharum spp.) is grown in muck and sand soils in Florida, which comprise of 78 and 22% of the hectarage, respectively. Field studies were conducted from 2004 to 2006 on muck and sand soils for one plant and two ratoon sugarcane crops in a cooperative genotype selection program based at Canal Point, Florida. The studies were conducted to determine the effects of muck and sand soils on the performance of sugarcane genotypes in the final stage (Stage IV) of testing in the breeding program in Florida. Data from six sets of 16 genotypes each from primarily the CP 04-, 05-, and 06- series and CPCL 95-, 02-, 05-, and 06- series in six trial locations on muck soils and three locations on sand soils were collected for three crop years (plant cane, first, and second ratoon). Analyses of mean squares on each soil separately, and for both soils combined, indicated that crop (plant cane, first, and second ratoon) made the highest contribution to the total mean squares. Crop × location was the interaction that made the most contribution to total mean squares. Tons of cane per hectare and tons of sugar per hectare showed crop × location interaction values of 2.95 and 4.38% on muck soils and 19.69 and 15.04% of the total mean squares on sand soils, respectively, which were higher than the other interaction components. Stability analyses indicated that only one genotype was stable for all characters in 2004 and 2005 on muck soils, while one each in 2005 and 2006 on sand soils. Similarly, only two genotypes were stable across both soil types. Usually genotypes that are stable are selected for release. These results indicate that it will be difficult for this program to consistently select genotypes with stable, high yields on both muck and sand soils. Separate tests with one set of genotypes selected from early stages on muck soils and a second set selected in early stages on sand soils are recommended for optimum genotype selection for each soil type. In these separate programs, Canal Point researchers should evaluate whether adding crop cycles to early selection stages and whether adding locations to the final selection stage on sand soils will improve selection decisions.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Sugarcane (Saccharum spp.) is grown in muck and sand soils in Florida, which comprise of 78 and 22% of the hectarage, respectively. Field studies were conducted from 2004 to 2006 on muck and sand soils for one plant and two ratoon sugarcane crops in a cooperative genotype selection program based at Canal Point, Florida. The studies were conducted to determine the effects of muck and sand soils on the performance of sugarcane genotypes in the final stage (Stage IV) of testing in the breeding program in Florida. Data from six sets of 16 genotypes each from primarily the CP 04-, 05-, and 06- series and CPCL 95-, 02-, 05-, and 06- series in six trial locations on muck soils and three locations on sand soils were collected for three crop years (plant cane, first, and second ratoon). Analyses of mean squares on each soil separately, and for both soils combined, indicated that crop (plant cane, first, and second ratoon) made the highest contribution to the total mean squares. Crop × location was the interaction that made the most contribution to total mean squares. Tons of cane per hectare and tons of sugar per hectare showed crop × location interaction values of 2.95 and 4.38% on muck soils and 19.69 and 15.04% of the total mean squares on sand soils, respectively, which were higher than the other interaction components. Stability analyses indicated that only one genotype was stable for all characters in 2004 and 2005 on muck soils, while one each in 2005 and 2006 on sand soils. Similarly, only two genotypes were stable across both soil types. Usually genotypes that are stable are selected for release. These results indicate that it will be difficult for this program to consistently select genotypes with stable, high yields on both muck and sand soils. Separate tests with one set of genotypes selected from early stages on muck soils and a second set selected in early stages on sand soils are recommended for optimum genotype selection for each soil type. In these separate programs, Canal Point researchers should evaluate whether adding crop cycles to early selection stages and whether adding locations to the final selection stage on sand soils will improve selection decisions.