全 文 ：Vol. 31 , No. 7
pp. 948 - 951 　July , 2005
作 　物 　学 　报
ACTA AGRONOMICA SINICA
第 31 卷 第 7 期
2005 年 7 月 　948～951 页
研究
简报
Analysis of Genetic Effects on Contents and Indexed of Protein and Oil in
Soybean Seeds in Different Environments
NING Hai2Long , LI Wen2Xia , LI Wen2Bin 3 , WANGJi2An
( Soybean Research Institute , Northeast Agricultural University , Harbin 150030 , Heilongjiang , China)
不同环境条件下大豆籽粒蛋白质和油分含量与指数的遗传效应分析
宁海龙　李文霞　李文滨 3 　王继安
(东北农业大学大豆研究所 ,黑龙江哈尔滨 150030)
　　Protein and oil are two of the important quantitative traits
closely related to the nutrient quality in soybean( Glycine max [L1 ]
Merry) 1 The content and yield of protein and oil have become one of
the main goals in soybean breeding1 For soybean breeding
programs , understanding the inheritance is of importance1 Genetic
analyses of protein and oil content in soybean seed have been
reported1 Most of the studies showed that protein and oil content are
quantitatively in2herited[1 - 9 ] 1 Genetic effects[1 - 3 ] ,
heritability[1 ,4 - 7 ] and correlation[8 ,9 ] of protein and oil content in
soybean seeds have been estimated1 In these experiments data were
analyzed mainly by analysis of average and variance ,which was not
available for that in different environments1 The contents of protein
and oil were unstable in different environments[10～15 ] 1 Zhu
integrated mixed model into plant genetic analysis and proposed
genetic models for quantitative traits in different environments[16 ] 1
Several studies have been conducted by using the model[17～19 ]1The
objectives of this study were to evaluate the additive and dominant
effects and their GE interaction effects ,to estimate the narrow2sense
heritabilities , to predict breeding merit of parents and estimate
correlation between quality traits for the content and index of protein
and oil in soybean seeds by this method1
1 　Materials and Methods
The mating design used for this experiment was an incomplete
diallel cross with 5 parents1 The materials , which were spring
soybean genotypes from Heilongjiang Province , comprised Dongnong
42 ,Dongnong 46 , Dongnong 7819 , Nongda 5129 , Heinong 35 , all
parents were crossed to each other in 20011 Some seeds of F0s were
planted in Hainan to obtain F1s seeds1 Seeds of parents ,F0s and F1s
were sown in the field at Northeast Agricultural University in 2002
and 20031 The experiments were laid out in a complete randomized
block design with three field replications1 There were 15 plants in
each plot for parents ,F0s and F1s at a spacing 65 cm×10 cm within
rows1 Seeds samples were taken at maturity from each of the 11 plants
in the middle part of each plot for the parent plants ,F1 and F21 The
quantitative traits of the soybean seed were protein content
(PC , %) ,protein index (milligrams of protein per soybean seed ,
PI) ,oil content (OC , %) ,oil index (milligrams of oil per soybean
seed ,OI) and the total content of protein and oil ( TC , %) which
were measured with three replications for each sample of parents ,F1s
and F2s1 PC and OC were determined in percentage on a dry matter
basis using near2infrared reflectance spectroscopy(NIR) and PI and
OI were gained by the product of PC and OC with seed weight1
Generation means were analyzed by genetic models for
quantitative traits of additive and dominant effects1 Genetic
components of additive variance ( VA ) , dominant variance ( VD ) ,
additive interaction variance ( VAE) ,dominance interaction variance
(VDE) were estimated by the MINQUE(1) method1 Genetic effect
was predicted by Adjusted Unbiased Prediction (AUP) method1
Estimates of variances and covariances were further used for
calculating heritability1 Breeding values of the parents were
evaluated according to magnitude of predictors1 The components of
genotypic correlation ( rG ) include additive correlation ( rA ) ,
dominance correlation ( rD) , additive interaction correlation ( rAE) ,
dominance interaction correlation ( r
DE
) 1 The Jackknife method was
used to derive the standard errors of estimated components of variance
and of predicted genetic effects by sampling the generation means of
genetic entries1 All data were analyzed by the programs of Zhu[16]1
Foundatinn item : National Natural Science Foundation of China (No1 30130120) ,National Brainstorm of Science and Technology of China (No1
2001BA511B0628) and Heilongjiang Province Education Commission (No1 10531003) 1
Biography : NING Hai2Long(1975 - ) ,male ,born in Zhaoyuan1 PhD of Agronomy1 Soybean Research Institute ,Northeast Agricultural University1 E2
mail : Ninghailong1975 @1631com1 3LI Wen2Bin :Corresponding author1 E2mail :wenbinli @yahoo1com
Received(收稿日期) : 2004208206 , Accepted(接受日期) : 20042122291

2 　Results
211 　Phenotypic value of parent and crosses for
protein and oil of soybean
　　Phenotypic value of parents ,F1 and F2 for PC ,PI ,OC ,OI and
TC are presented in table 11 Difference between parents or crosses indicated the possibility of improving soybean for protein and oiltraits and studying the genetic mechanism involved1 In F1 ,the meansof PC , PI and TC were much lower than those of parents , while therelation of mean of F1 and parent for OI was different in 2 years1 InF2 ,the mean of PC , PI and TC were much lower than F1 ,while themean of OC and OI were larger than that of F11
Table 1 Phenotypic mean values ( range) of generations for PC, PI,OC,OI and TC
Generation Year PC PI OC OI TC
Parents
2002
2003
421670 ±11822 131078 ±11266 181702 ±11703 51359 ±01727 611372 ±01598
(401880～451620) (111703～141441) (161520～201860) (41681～61460) (601770～621140)
421396 ±21377 121477 ±11253 181872 ±11659 51148 ±01250 611268 ±11001
(391400～441810) (101730～131808) (171140～201890) (41916～51511) (601290～621590)
F1
2002
2003
411059 ±11211 121036 ±01818 191811 ±11028 51416 ±01491 601870 ±01447
(391440～421900) (101595～131593) (181270～211100) (41617～61277) (601540～611930)
411559 ±11210 111627 ±11193 191624 ±01995 51118 ±01538 611283 ±01452
(391850～441050) (91818～141058) (181160～211270) (41417～51887) (601550～621210)
F2
2002
2003
401589 ±01964 111428 ±01865 201828 ±01642 51471 ±01575 611426 ±11138
(391730～421030) (101363～121379) (191730～211400) (41632～61802) (601620～631430)
401560 ±01810 111562 ±01412 201660 ±01429 51491 ±01375 611422 ±01488
(391820～411460) (111233～121146) (201210～211270) (51107～51911) (601660～641720)
212 　Estimation of genetic variance
According to the magnitude of each genotypic variance component ,
PC , PI , OC , OI and TC of soybean were affected by the major
contribution effects1 Estimates of the variances components of PC ,PI ,
OC ,OI and TC of soybean are summarized in table 21 The results
indicated that the traits of PC ,PI ,OC ,OI and TC were mainly affected
by genetic main effects of genes and the genetic variance (VG = VA +
VD) account for 100 % , 68 % , 67 % , 100 % and 100 % of the total
genetic variance[VG + VGE (VGE = VAE + VDE) ] ,respectively1
Protein content of soybean seed was controlled by additive effects
(VA)1 For PI ,additive effect (VD) and additive interaction effect (VAE)
were significant at the 0101 probability level1 The OC was mainly
controlled by additive effects ( VA ) ,followed by additive interaction
effects(VAE)1 The OI and TC were controlled by additive effects (VA)
and dominant effects (VD)1
Table 2 Estimation of genetic variances and
covariances of PC, PI,OC,OI and TC
Variance
components PC PI OC OI TC
VA 11546 3 3 01000 11098 3 3 01204 3 3 01294 3 3
VD 01000 11857 3 3 01000 01200 3 3 01302 3 3
VAE 01000 01886 3 3 01538 3 3 01000 01000
VDE 01000 01000 01000 01000 01000
　　Note : 3 3 means significant at 0101 level .
213 　Estimation of heritabilities
Estimates of the narrow2sense heritability for PC , PI ,OC ,OI
and TC are presented in table 31 The total narrow2sense heritabilities
for PC , PI ,OC ,OI and TC were 2814 % ,1916 % ,3012 % ,2018 %
and 2110 % , respectively1 For PC , OC , OI and TC , the general
heritabilities were 2814 % , 2115 % , 2018 % and 2110 % ,
respectively1 Interaction heritabilities for PI and OC were 1916 %
and 1015 % , respectively1 Selective advantage for PC ,OC ,OI and
TC could be expected in the early generations1 The expression of
genes for protein index and oil content were not stable in different
environments ,so the selection would not be efficient1
Table 3 Estimation of heritabilities
of PC, PI,OC,OI and TC in soybean seed
Heritability PC PI OC OI TC
h2N 01284 3 3 01000 01215 3 3 01208 3 3 01210 3 3
h2N E 01000 01196 3 3 01105 3 01000 01000
　　Note : 3 3 , 3 mean significant at 0101 and 0105 probability levels ,
respectively1
214 　Prediction of additive effect of parents
The additive effects with significant variance in table 2 were
used to predict the merit values of the parents used in this experiment
(Table 4) 1 Dongnong 42 and Heinong 35 were better than other pa2
rents for an improved protein content and total content of protein and
oil1 Dongnong 7819 was a better parents for improving oil content
and oil index1
Table 4 Predicted additive effects
of parents for PC,OC,OI and TC
PC OC OI TC
Dongnong 42 01432 3 - 01770 3 - 01243 3 - 01224 +
Dongnong 46 - 01809 01535 01169 - 01241
Dongnong 7819 - 01554 01521 + 01364 3 - 01086
Nongda 5129 01091 01109 - 01047 01150
Heinong 35 01839 3 - 01395 - 01243 01400 +
　　Note : 3 , + mean significant at 0105 and 0110 probability level ,
respectively1
215 　Estimation of correlation in soybean seed
Since the total genetic effects could be further partitioned into
949　第 7 期 NING Hai2Long , et al1 :Analysis of Genetic Effects on Contents and Indexed of Protein and Oil in Soybean··· 　　　

genotypic effects and GE interaction effects , the correlations
consisted of genotypic correlation and GE interaction correlation1
Additive correlation ( rA ) , dominance correlation ( rD ) , additive
interaction correlation ( rAE) and dominance interaction correlation
( rDE) were presented in table 51 Additive correlations ( rA ) for PC
and OC , PC and OI ,OI and TC were significantly negative ,but for
PC and TC , OC and OI were positive1 It was concluded that
simultaneously improving PC and OC , PC and OI , OI and TC is
difficult , while improving PC and TC , OC and OI is possible in
soybean quality breeding1 Dominance correlations ( rC) for PI and
OI ,PI and TC were significantly positive1 It suggested that improving
these pairwise traits simultaneously were relatively difficult1 Additive
interaction correlation ( rAE)for PI and OC was negative1 It indicated
that efficiency of improving PI and OC would be small in different
environments1
Table 5 Estimation of correlation of genetic components of PC, PI,OC,OI and TC in soybean seeds
Correlation PC and PI PC and OC PC and OI PC and TC PI and OC PI and OI PI and TC OC and OI OC and TC OI and TC
rA 01000 - 01196 3 - 01483 3 3 01203 3 01000 01000 01000 01515 3 3 - 01081 - 01185 3
rD 01000 01000 01000 01000 01000 01372 3 3 01195 3 3 01000 01000 01112
rAE 01000 01000 01000 01000 - 01162 3 01000 01000 01000 01000 01000
rDE 01000 01000 01000 01000 01000 01000 01000 01000 01000 01000
3 　Discussion
Inheritance of soybean quality traits has been widely studied in
recent several decades1 Weber determined protein and oil content in
an interspecific cross and concluded that oil content was controlled
by additive effect and protein content by additive and dominant
effects[1 ]1 Chuanhan et al1[2 ] and Chen et al1[3 ] considered that
protein and oil content were controlled by additive and dominant
effects by analysis of combination ability1 In the present study , PC
and OC was controlled by additive effects which was similar with
prior reports1 But the estimates of dominant effects about PC and OC
were zero ,which was contrary to prior reports1 The reason might
come from different parents used for the crosses or misunderstanding
with other genetic effects1 It is necessary to further study1 PI was
controlled by dominant effect , and OI and TC were controlled by
additive and dominant effects1
The contents of protein and oil were influenced by environments ,
which means the two traits were unstable in different environments1
Wang called the contents of protein and oil as ecological traits and
described the performance of the traits in different environments[10] 1
Erickson et al1[15] ,Ning et al1[11 ,12] and Zhang et al1[13 ,14] found
important genotype ×environment interaction effects on the contents of
protein and oil1 But the genetic mechanism of the genotype ×
environment interaction effects have not been studied1 In this
experiment , by using the genetic model and statistical method[16] ,
genotype ×environment interaction effects were analyzed , indicating that
protein index and oil content were controlled by seed additive interaction
variance ,and the improvement of quality in soybean seeds should be
conducted in special environment1
Heritability is an important parameter for soybean breeder to
adjust breeding program1 The trait with a higher heritability can be
selected in early generations ,while the traits with lower heritability
should be selected in late generation1 Weber et al1 reported the heri2
tability of protein content in F2 was 70 %[1 ] 1 Mekendry et al1
estimated the heritability in F2 was 61 % - 72 %[4 ]1 Johnson et al1
reported the heritability in F3 family lines was 39 % - 83 % ,after F5 ,
the average of heritability reached 81 % - 95 %[5 ] 1 Weber et al1
found the heritabilities of OC in single soybean plant in F2 from 3
crosses were 49 % ,56 % and 59 % ,respectively[6 ] 1 Hanson et al1
found the heritabilities of OC in F2 ,F3 and F4 were 34 % ,61 % and
62 % ,respectively[7 ] 1 In the present experiment , the total narrow2
sense heritability for protein content ,protein index ,oil content ,oil
index and total contents of protein and oil were 2814 % , 1916 % ,
3012 % ,2018 % and 2110 % , respectively1 For protein content ,oil
content ,oil index and total contents of protein and oil , the general
heritability values were 2814 % , 2115 % , 2018 % and 2110 % ,
respectively1 Interaction heritability values for protein index and oil
content were 1916 % and 1015 % ,respectively1 Selective advantage
for protein content , oil content ,oil index and total contents of protein
and oil could be expected in the early generations1 The expression of
genes for protein index and oil content was not stable in different
environments ,so the selection would not be efficient1
According to the magnitude of genetic main effects ,the genetic
patterns of parents can be further illustrated for soybean quality
traits ,which are of importance in improving soybean quality through
selection of parents1 In the present study ,it was showed that some
parents had better breeding values due to genetic main effect1
Dongnong 42 and Heinong 35 were better parents for improving
protein content and total contents of protein and oil and Dongnong
7819 was a better parent for improving oil content and oil index1
During the course of soybean quality improving , enough
emphasis should be given on the correlation of PC and OC1 Many
experi2ments gave the similar conclusion that there were significant
negative correlation between PC and OC[8 ,9 ]1 In the present study ,
correlations were further partitioned into additive correlation ,
dominance correlation ,additive interaction correlation and dominance
059 　　　　作 　　物 　　学 　　报 第 31 卷 　

interaction correlation1 Additive correlation can be inherited to next
generations and be efficient to guide breeding1 The results showed
that negative additive correlation was highly significant , suggesting
that it is difficult to improve PC and OC simultaneously1 Negative
additive correlation for PC and OI , OI and TC resulted in the
conclusion that simultaneously improving PC and OC ,PC and OI ,OI
and TC are difficult ,but positive additive correlation for PC and TC ,
OC and OI show a feasibility for improving PC and TC ,OC and OI in
the same time1 Dominance correlations are unstable among different
generations and it is unreliable for breeding program , suggesting that
improving PI and OI , PI and TC simultaneously are relatively
difficult1 Significant additive interaction correlation for PI and OC
indicated that efficiency of improving PI and OC would be small in
different environments1
References
[1 ] 　Weber C R1 Inheritance and interrelation of some agronomic and
chemical characters in an interspecific cross in soybeans G1 max ×G1
ussuriensis1 Towa Agricultral Experiment Research Bulletein ,1950 ,374 :
767 - 816 　　　
[2 ] 　Chuanhan V S ,Singh B B1 Genetic analysis of protein and oil content in
soybean1 India Journal of Agricultural Science ,1983 ,53 :634 - 637
[3 ] 　Chen H2H(陈恒鹤) ,Yin L2H(尹丽华) ,Wang D2Q(王大秋) ,Fu Y2
H(傅艳华) ,Li N(李楠) ,Yang Y2H(杨玉环) ,Liang Z2F(梁振富)1 Study on inheritance and selection effect of protein and oil content in
soybeans 31 variability and genetic advance in early generations1
Soybean Science (大豆科学) , 1991 ,10 (1) : 1 - 9 ( in Chinese with
English abstract)
[4 ] 　Mekendry A L1 Inheritance of seed protein and seed oil content in early
maturing soybean1 Candia Journal of Genetic Cytology ,1985 ,27 :603
- 607
[5 ] 　Johnson H W , Robison H A ,Comstock R E1 Estimates of genetic and
environmental variability in soybean1 Agronomy Journal ,1955 ,47 :314 -
318 　　
[6 ] 　Weber C R ,Moorthy B R1 Heritable and nonheritable relationship and
variability of oil content and agronomic characters in the F2 generation
of soybean crosses1 Agronomy Journal ,1952 ,44 :202 - 209
[7 ] 　Hanson W D1 Analysis of genetic variability from generation of plant
progeny lines in soybeans1 Crop Science ,1962 ,2 :63 - 67
[8 ] 　Johnson H W ,Robison H A ,Comstock R E1 Genotypic and phenotypic
correlations in soybean and their implications in selection1 Agronomy
Journal ,1955 ,47 :477 - 483
[9 ] 　Burton J W1 Breeding soybean for improved protein quantity and
quality1 WSRC Ⅲ,19851 361 - 367
[10 ] 　Wang J2L (王金陵) 1 Ecological Type of Soybean(大豆生态类型) 1
Beijing : China Agriculture Press ,19911 30 - 74(in Chinese)
[11 ] 　Ning H2L (宁海龙) ,Yang Q2K(杨庆凯) ,Pan X2W(潘相文) ,Wang
H2X(王红霞) 1 Analysis of protein and fat content of soybean cultivars
recently tested in Heilongjiang Province1 Journal of Northeast
Agricultural University (东北农业大学学报) , 2002 ,33 (4) :319 -
323(in Chinese with English abstract)
[12 ] 　Ning H2L (宁海龙) , Zhang D2Y(张大勇) , Zhang S2Z(张淑珍) ,
Yang Q2K(杨庆凯) 1 Ecological features of oil and protein content of
soybean in Northeast China1 Soybean Science (大豆科学) , 2003 ,22
(2) :132 - 136(in Chinese with English abstract)
[13 ] 　Zhang D2Y(张大勇) ,Ning H2L (宁海龙) , Yang Q2K(杨庆凯) 1
Difference of oil and protein contents of several soybean varieties to the
ecological condition in northeast of China1 Chinese Journal of Oil Crop
Sciences (中国油料作物学报) ,2003 ,25 (1) :18 - 20 ,24(in Chinese
with English abstract)
[14 ] 　Zhang D2Y(张大勇) ,Ning H2L (宁海龙) , Hu G2H (胡国华) ,
Zhang S2Z(张淑珍) ,Wang J2A (王继安) ,Li W2B (李文滨) 1 The
analysis of the effect on sites and years to soybean protein and oil con2
tent in Northeast China1 Soybean Science (大豆科学) ,2004 ,23 (1) :
30 - 35(in Chinese with English abstract)
[15 ] 　Erickson ,L R ,Beversdorf W D ,Ball S T1 Genotype ×environment
interactions for protein in Glycine max ×Glycine soja crosses1 Crop
Science ,1982 ,22 :1 099 - 1 101
[16 ] 　Zhu J (朱军) 1 Analysis Methods for Genetic Models(遗传模型分析
方法) 1 Beijing : China Agriculture Press , 1997(in Chinese)
[17 ] 　Shi C H ,Zhu J ,Zang R C ,Chen GL1 Gentic and heterosis analysis for
cooking quantity traits of indica rice1 Theor Appl Genet ,1997 ,95 :294 -
300 　　　
[18 ] 　Shi C H , Xue J M , Yu , Y G, Yang X E , Zhu J1 Analysis of genetic
effects on nutrient quality traits in indica rice in different environment1
Theor Appl Genet ,1996 ,92 :1 099 - 1 102
[19 ] 　Wu J2X(吴吉祥) ,Wang G2J (王广金) ,Zhu J (朱军) ,Xu F2H(许
馥华) ,Ji D2P (季道藩) 1 Genetic analysis on direct and maternal
effects of seed traits in upland cotton( Gossypium hirsutum L1) 1 Acta
Agronomica Sinica (作物学报) ,1995 ,21 (6) :659 - 664 (in Chinese
with English abstract)
159　第 7 期 NING Hai2Long , et al1 :Analysis of Genetic Effects on Contents and Indexed of Protein and Oil in Soybean···