Document Type: Research Paper
Department of Plant Science and Agricultural Resources, Faculty of Agriculture, Khon Kaen University, Khon
Kaen, 40002, Thailand.
Multi-environment trials (METs) of crop genotypes are costly and require
efficient test sites for cost effectiveness. This study aimed to identify efficient
test sites for METs of sugarcane (Saccharum spp.) genotypes in Thailand, utilizing
data from 10 sugarcane genotypes conducted at nine locations covering different
sugarcane growing regions of the country for two crop-classes. Cluster analysis and
the genotype plus genotype × environment (GGE) biplot method were used to group
these sites into five subsets, based on their similarity in genotypic responses of cane
and sugar yields of the planted crop and the first ratoon crop. The results showed a
fair agreement between the two methods, but inconsistent results were obtained
from groupings that were based on different yield traits and crop-classes. Locations
appearing more consistent in certain groups were chosen as the representatives of the
respective groups to constitute the set of efficient test sites. Cluster analysis and the
GGE biplot, however, identified different sets of test sites that were equally effective
in retaining the G×L interaction and the performance ranking of the test genotypes as
the original nine test sites. The selected locations by cluster analysis which included
Nakhon Ratchasima, Ratchaburi, Kamphaeng Phet, Tha Phra, Khon Kaen and Udon
Thani are preferred because of their wider geographical distribution. Four sites could
thus be omitted, which would substantially reduce the costs and time and greatly
improve the efficiency of the METs of sugarcane genotypes in Thailand.
Keywords: Multi-environment trials (METs); Environment grouping; GGE biplot;
Cluster analysis; Breeding line evaluation.