|
| |
REGULAR ARTICLE
Pattern analysis of multi-environment trials in bread wheat
Yuksel Kaya, Mevlut Akcura, Ramazan Ayranci, Seyfi
Taner
Communications in Biometry and Crop Science (2006) 1 (2), 63-71.
ABSTRACT
To detect genotype-by-environment interactions (GEIs), pattern analysis
(PA) was performed on yield data of 20 bread wheat (Triticum aestivum
L.) genotypes tested across nine environments during the 2001-2002
growing season in the Central Anatolian Region of Turkey. Nine clusters
of genotypes with similar patterns in performance, mostly reflecting
their origin and pedigree, were identified. Most of the genotypes from
the National Bread Wheat Improvement Program (NBWIP), Turkey, fell into
one of the two broad genotypic clusters, whereas most of those from the
International Winter Wheat Improvement Program (IWWIP), a collaborative
breeding program among Turkey, CIMMYT and ICARDA, were clustered
together in another cluster. Six clusters of the environments with
similar trends in discriminating genotypes were identified;
discrimination of the environments tested indicated that there existed
two distinct main types of environments: rain-fed and irrigated. However,
the rain-fed environment E2 (Eskisehir) joined the irrigated
environments, whereas the irrigated environment E9 (Haymana) was
clustered with the rain-fed environments. Irrigated environments E6 (Konya)
and E7 (Cumra) contributed trivially to discrimination of the genotypes,
as they exhibited a pattern that was almost identical to that of the
irrigated environments. The first two principal components explained
53.3% of the total variation in GEI data. This study revealed that the
genotypes with a common parent in their pedigree or with the same origin
tended to be clustered together. All the genotypes from the NBWIP and
one-third of genotypes from the IWWIP were stable; therefore, the NBWIP
could be strengthened by introductions from the IWWIP. Environmental
cluster analysis effectively identified trials that received rainfall
and supplementary irrigation. It might be possible to reduce the number
of test environments by eliminating one or more of those that
differentiate among genotypes in a similar manner (e.g., irrigated
environments E6 in Konya and E7 in Cumra).
Key Words: bread wheat (T. aestivum
L.); cluster analysis; multi-environment trials; genotype-by-environment
interaction; pattern analysis; principal component analysis.
|
