|
18.
|
Study on genetic effects for lysine traits of indica rice
in different environments
|
|
C.
|
Shi, J. ZHU, Y. Yu, X. Yang and J. Wu
|
Thejiang Agricultural University, Hangzhou, 310029 China
Lysine in rice (Oryza sativa L.) is of importance for health
of people whose main food in daily life is rice. As an important part of
seed, the diploid embryo might have genetic effects affecting the performance
of rice quality traits as well as endosperm, cytoplasmic and maternal genetic
effects (Shi et a!. 1995). But there was litter information about the expression
of diploid embryo nuclear genes for the performance of lysine traits in
rice. The objectives of this study were to evaluate the diploid embryo,
triploid endosperm, cytoplasmic and diploid maternal plant genetic effects,
as well as their genotype x environment (GE) interaction effects; and to
estimate components of variance and heritability for lysine traits of indica
rice.
Nine cytoplasmic male sterile lines (Thexie 2 A, etc.) or
their maintainer lines (B) and five restorer (R) lines (T 49, etc.) as
parents were used in an incomplete diallel cross in two years. Seedlings
of parents, F1 (A x R) and RF1 (R x B) were planted in the field of experimental
farm at Thejiang Agricultural University in 1995 and 1996. The Seeds were
sown on April 2 in both years and single plants of 30-day old seedlings
were transplanted at spacing of 20 x 20 cm. There were 24 plants in each
plot with three replications. Seed samples of parents, F2 (A x R) and RF2
(R x B) were harvested at maturity from 8 plants in the middle part of
each plot. The F1 seeds from A x R and BC1 seeds from A x F1 were obtained
by crossing females to males during the flowering season. Quantitative
traits analyzed were lysine content (LC, %), Lysine index (LI, mg lysine
per milled rice) and the ratio of lysine content to protein content (RLP)
of milled rice, which were measured with three replication for each sample
of parents, F1, F2 , RF2 and BC1.
The genetic main effects and their GE interaction effects
of diploid embryo, triploid endosperm, cytoplasmic and diploid maternal
plant were analyzed for rice lysine traits by using the full genetic model
for quantitative traits of seed in cereal crops (Zhu 1997). According to
the model, the phenotypic variance (Vt) can be further partitioned into
several components. Partitioning for the phenotypic variance (Vt) is:
Vp= Vg + VGE + Ve
|
=
|
VA0 + VDo + VAe + VDe + VC ÷ VAm + VDm
|
|
+
|
VAOE + VDOE + VACE + VDeE + VCE + VAmE + VDmE + Ve
|
where VG = genetic main variance, VGE = GE interaction variance,
VA0 = embryo additive variance, VD0 = embryo dominance variance, VAe =
endosperm additive variance, VDe = endosperm dominance variance, Vc = cytoplasm
variance, VArn = maternal additive variance, VDm = maternal dominance variance,
VAOE = embryo additive interaction variance, VDOE= embryo dominance interaction
vanaflce, VAeE= endosperm additive interaction variance, VDeE= endosperm
dominance interaction variance, VCE= cytoplasm interaction variance, VAmE
= maternal additive interaction variance, VDmE= maternal dominance interaction
variance, and VE= residual variance.
Since the total genetic effect can be partitioned into components
for genetic main
49.51%, respectively). In genetic main effects, LC was mostly
contributed by cytoplasmic main effects, since cytoplasmic main variance
(Vs) was larger than embryo main variance (VAO and Vdo), endosperm main
variance (VAe and Vde) or maternal main variance (Vam and Vdm). While LI
and RLP were mainly contributed by embryo and endosperm main effects, respectively,
since embryo main variance (VAO and Vdo) and endosperm main variance (VAe
and Vde) was larger than other genetic variances. For GE interaction effects,
embryo interaction effects were much more important than other interaction
effects for LC and RLP because of the larger embryo interaction variance
(VAOE + VDOE), and maternal interaction effects for LI because of the larger
maternal interaction variance (VAmE + Vdme). Since the embryo interaction
effects and maternal interaction effects were more important in genetic
interaction effects for lysine traits, the expression of diploid embryo
genes and maternal plant genes was more easily affected by environments
than that of triploid endosperm genes and cytoplasm genes.
The estimates of total narrow-sense heritability (h2) were
57.60, 59.47 and 70.87% for LC, LI and RLP, respectively (Table 2). The
general heritabilities (h2g) controlled by additive effects in genetic
main effects were 41.00, 52.46 and 6.05% for LC, LI and RLP, respectively.
Therefore,the general heritabilities were more important for LC and LI,
and the selection for these two quality traits was stable in different
environments. The interaction heritabilities (h2gE) controlled by additive
interaction effects were 16.60, 7.01 and 64.82% for LC, LI and RLP, respectively.
So the interaction heritability was more important for RLP and the selective
effect for RLP is easily affected by environments. For the general heritability,
cytoplasm general heritability (h2c = 23.89%), embryo heritability (h2go
= 29.01%) and endosperm general heritability (h2ge = 5.56%) were more important
than other general heritability components for LC, LI and RLP, respectively,
while embryo interaction heritability for LC (h2goE = 16.60%) and RLP (h2goE
45.3 1%) or cytoplasmic interaction heritability (h2cE = 7.01%) were more
important for these quality traits in interaction heritability.