Summary of agricultural dissertation: Agronomic traits associated with drought tolerance of tropical-derived germplasm for hybrid maize breeding

MINISTRY OF EDUCATION AND TRAININ

MINISTRY OF AGRICULTURE AND RURAL DEVELOPMENT

VIETNAM ACADEMY OF AGRICULTURAL SCIENCES
---------------------------------

DO VAN DUNG

AGRONOMIC TRAITS ASSOCIATED WITH DROUGHT
TOLERANCE OF TROPICAL-DERIVED GERMPLASM
FOR HYBRID MAIZE BREEDING

Specialized: Genetics and Plant Breeding
Code: 9.62.01.11

SUMMARY OF AGRICULTURAL DISSERTATION

HANOI - 2018


The work was completed in:
VIETNAM ACADEMY OF AGRICULTURAL SCIENCES

Science advisor:
1. Ph.D. Le Quy Kha, Institute Of Agricultural Science for Southern
Vietnam
2. Ph.D. Pervez Haider Zaidi, International Maize and Wheat
Improvement Center (CIMMYT) in India.

The dissertation would be defended on Foundation level examination board

Vietnam Academy of Agricultural Sciences
August, 2018

Dissertation can find out at:
1. Vietnam Academy of Agricultural Sciences
2. Maize Research Institute


1
GENERAL INTRODUCTION
In Vietnam, drought is one of major constraints for maize production.
Consequently, damage caused by drought was estimated more than 30%, up to 7080% or no harvest at all. Vietnam is one of the most affected countries by climate
change and drought has been more and more regular and severe. Therefore, it is
annually about 0.6 and 0.7 million ha maize area facing more stresses especially
drought. Maize area, grain yield and production in Vietnam in 2017 was 4.6
tons.ha-1 respectively that it is lower than the global average (5.5 tons.ha-1).
Meanwhile, the demand becomes more and more, leading to exceeding the supply
Therefore, breeding drought tolerant maize varieties is the most important
orientation for rainfed maize areas. Therefore, it is essential for the thesis
“Agronomic traits associated with drought tolerance of tropical-derived
germplasm for hybrid maize breeding” for increasing an efficiency in breeding
gain may be crucial.
RESEARCH OBJECTIVES
Studies on agronomic traits associated with drought tolerant ability of some
materials were carried out for selection of inbred lines and development of
promising maize hybrids for production.
SCIENTIFIC AND PRACTICAL MEANINGS OF THE THESIS
- The meanings of science
Providing more and more scientific data on phenotyping and evaluating combining

ability of F2:3 BP progenies populations under managed drought and optimal
condition, in combination with identifying genome regions associated with
quantitative trait locus (QTLs) controlling drought tolerance at the early generations
during improvement and development of drought tolerant maize varieties.
- The meanings of practice
+ Based on the results of evaluating drought tolerance of 8 BP progenies at F2:3
generations [8 BP populations × testers] for selection and development of 9 inbred
lines with good combing ability and drought tolerance, high yield under drought
stress, higher efficiency in drought tolerant maize breeding.
+ Having identified 2 significant gene clusters, including the first on
chromosomes 1 (bin 1.05-1.07), 7 (bin 7.01-7.03) and the other on chromosome 8
(bin 8.02-8.03), controlling anthesis- ilking interval, leaf senescence and grain
yield relating to drought tolerance.


2
+ Having developed 9 inbred lines (RA1, RA2, RA3 ... RA9) used as drought
tolerant materials for breeding maize toward rainfed condition.
+ Having developed 2 promising hybrid combinations (LVN72, ÐH17-1)
suitable for rainfed condition in Vietnam.
NEW CONTRIBUTIONS OF THE THESIS
Giving more data on phenotypes and quantitative trait loci (QTLs) on genome
regions of chromosomes 1, 4, 6 7 and 8 associated with drought tolerance for
development of germplasm and maize breeding for rainfed condition. Having
developed 9 inbred lines with good combining ability, drought tolerance, high grain
yield and initially introduced 2 promising hybrid combinations (LVN72, ÐH17-1)
for production.
OBJECTS AND SCOPE OF RESEARCH
Researches on 8 F2:3 populations (790 families), crossed from 10 CIMMYT
derived tropical inbred lines and test-crosses (2 testers called CML451, CLO2450),

through which 9 best families were selected to develop into 9 inbred lines, crossed
into 36 crosses by diallel. Local checks: in India these were PAC754, 30V92,
HTMH5401 and 900MG; while using LVN10, VN8960, LVN61, NK67, C919,
DK9901 in Vietnam. Experiments were carried out under managed drought and
optimal conditions at International in Hyderabad, India and Ninh Thuan Province,
Vietnam; Testing hybrid combinations were done in the northern region of Vietnam.
ORGANIZATION OF THE THESIS
The thesis contains 138 pages, 32 tables, 15 pictures and graphs with 5 sections
general introduction (4 pages), chapter 1: Literature review (44 pages), chapter 2:
Materials and methods (13 pages), chapter 3: Results and discussion (77 pages), and
Conclusion and suggestion (2 pages); References of 201 documents with 29
Vietnamese ones and 169 others in English and 3 cited from websites; in which 2
research works published on Journal of Vietnam Agricultural Science and
Technology and the other at the 12th Asian Maize conference in Bangkok, Thailand.
Chapter 1. LITERATURE REVIEW
1.1. The global maize production situations and in Vietnam
1.1.1. World maize production
Maize (Zea mays L.) have been of profound changes. Maize area, grain yield
and production in 2010 higher increase than in 2005 by 9.8% area, 7.8%


3
productivity and 18.3% production. And there was, in 2015, an increase of 9.5%
area, 5.27% yield and 15.3% production. In 2017/2018, the forecast of maize
production will be 1,046 million tons, 42 million tons lower than in 2016/2017
(1,088 million ton). These data show that an increase in maize production has been
slower for recent years but in the future, more and more will demands on maize be,
mainly for feed and by 2050 doubled with 1,178 million tons and 194 million
hectares. In fact, world maize areas, grain yield and production develop drastically
but mainly in developing countries where maize cultivation still depends on rainfed

conditions. Therefore, it is necessary to continuously improve drought tolerance of
maize varieties for an increase in grain yield and production.
1.1.2. Maize production in Vietnam
Maize production were obtained from 1995 to 2004, with annual increase in
areas (5.3%), yield (4.8%) and production (10.7%) respectively. During 20052015, domestic maize production still went on the same trend but more slowly that
it was annually only 2.2% in grain yield, 2.0% in area và 5.0% in production. On
the other hand, for more than 10 last years, the livestock gets an average increase
of 8 - 12% per year, so maize must be imported for domestic demand and it was
about 8,8 million tons in 2017 and as predicted, 10.5 million tons in 2018.
Therefore, it is necessary to continuously improve grain yield, quality and drought
tolerance of maize varieties.
1.2. Impacts of drought on maize production in the world and Vietnam
1.2.1. Impacts on global maize production
More and more severely and unpredictably has climate change been taking
place worldwide, in which drought can be considered as one of main factors.
Annually, the loss of maize production due to drought is around 8%. By 2025,
drought will have got more severe and regions will have become drier in all over
the world but mainly in Africa and Asia. Impacts due to climate change, at present
and in the future, can globally affect nearly 160 million hectares and reduce maize
production by 6 - 23%. Therefore, in fact, it is essential to develop drought tolerant
maize varieties remain stability and increase in production, satisfactory to more
and more demands.
1.2.2. Impacts on maize production in Vietnam
In Vietnam, drought is a major constraint for maize production, there is about


4
0.3 million ha facing the shortage of water and the loss of 0.5-0.7 million tons.
Though, drought takes place in all 8 maize regions, in which the most severe
drought stress in North Central Region, South Central Coast Region and Central

Highland Region; intermediate stress in Mountainous Region, South East Region
and the Mekong River Delta Region; moderate stress in the Red River Delta
Region. Otherwise, global total surface water in 2025 will be about 96% of in 2010,
by 2030 total surface water inflows at the upstream, deltas and the Red river valleys
will have been reduced by 2.4%; 2.9% and 1.9%, respectively and water shortage
will become serious in 50 years. Hence, breeding maize with high yield, drought
tolerance and stability is really crucial for enhancement of maize production.
1.3. Research and usage of drought tolerant maize varieties
1.3.1. Research and usage of drought tolerant maize in the world
Over the past 38 years, breeders have been selecting and improving drought
tolerance in maize. The results of having developed drought tolerant maize since
2008 were summarized: breeding gain with traditional methods is 50 kg.ha-1
(equivalent to 1.4% per year) while with marker assisted selection- MAS and gene
transfer methods are 20 kg.ha-1 (equivalent to 0.6%) and 30 50 kg.ha-1 (equivalent
to 0.7%) respectively. Studying characteristics of active genes or molecular
markers closely related to genes controlling drought tolerance is an important step
in applying the selection of genotypes in improving drought tolerance in maize.
1.3.2. Researches of drought tolerant maize in Vietnam
Since 1990s, there has been of full researches on drought tolerance at the stages
of maize (from the seedling to flowering); During 1988-1998, studies were focused
on high plant density, anthesis-silking interval (ASI), leaf senescence, grain yield and
ect; Recently, applying biotechnology in maize breeding, in which, mainly focusing
on two fields: tissue culture and recombinant DNA technology for an improvement
in grain yield has been developed. Identifying molecular markers relating to drought
tolerance such as Dhn gene helped select drought tolerant germplasm exactly. Hence,
it may be concluded that combining conventional and modern approaches is a basis
for breeding maize varieties tolerant to drought in Vietnam.
1.4. Scientific basis of drought and drought tolerance in maize
1.4.1. Definitions of drought
Drought is a harsh condition and the consequence of the shortage of rainfall or no

rain for prolong more than crop season or no enough water. Drought is classified as


5
following: In low land tropics a marginal rainfed maize environment may be defined
as having seasonal precipitation below 500mm; in highlands seasonal precipitation
below 350 mm; During bracketing- flowering stage, less than 100mm rainfall is
considered as drought, between 100-200 mm as being marginal for maize production.
1.4.2. Maize under drought environment
Though, every crop stage of maize has some susceptibility to drought, however,
three stages of early growth stage (when plant stand are established), flowering and
mid-to-late grain filling stage, are considered critical stages to drought, especially at
flowering. Drought leads to reducing the growth of leaf, silk, stem, root, grain yield.
1.5. Genetics of drought tolerance in maize
Drought tolerance in maize is controlled by multiple genes (multigene) and
their environment. Maize can express drought tolerant ability in many ways such
as drought avoidance and tolerance during growth and development in order to
reduce grain yield loss. The heritability, the correlation between parents and
generations, helps predict heterosis based on the correlation of secondary traits and
grain yield. The importance of secondary traits for breeding maize for drought
tolerance is their correlation with grain yield, in which, it should be more focus on
traits of anthesis-silking interval, ear per plant, leaf senescence and ect.
1.6. Useful traits for breeding maize tolerant to drought
Some secondary traits are used for maize breeding for drought tolerance: 1)
Leaf rolling at crop stage of seedling to pre-flowering; 2) root system; 3) Ears per
plant; 4) Anthesis-silking interval (ASI); 5) Leaf senescence (SEN); 6) stay- green;
7) Shelling percentage; 8) efficient ear length; 9) Grain yield.
1.7. Application of molecular assisted selection
1.7.1. Application of molecular assisted selection for maize breeding
Since the beginning of the 20th century, marker assisted selection (MAS) has been a

useful tool in maize development and improvement. This approach enables breeding
based on genotypes, so markers associated with one or many genes controlling
interested traits through which it is able to identify germplasm with stress tolerant
genes. Now, applying quantitative trait loci (QTLs) mapping on genome regions of
chromosomes helps exactly identify anticipated materials and save time.
1.7.2. Single nucleotide polymorphism (SNP)
Single nucleotide polymorphism (SNP) called as "snips", is the most popular
type of genetic variation. Each SNP represents a variation in a single nucleotide


6
that occurs at a specific position in the genome, where each variation is present to
some appreciable degree within a population (e.g.> 1%).
1.7.3. QTL Mapping and genetics of quantitative traits
A quantitative trait locus (QTL) is a region of DNA which is associated with a
particular phenotypic trait, which varies in degree and which can be attributed to
polygenic effects, i.e., the product of two or more genes, and their environment.
To define genes controlling traits should be based on the combination of genotypic
and phenotypic analysis at segregating populations while mapping QTLs based on
mathematical models.
1.7.4. Improvement by conventional approaches and QTL mapping
By effectively using genetic diversity, developing various elite lines, particularly
through recurrent selection for interested genotypes in heterozygous and recombinant
populations (F2, F2:3). The combination of traditional methods, phenotypic
assessments under different environments and assistance of advanced biotechnology
tools (such as SNPs) showed, through each selection cycle, grain breeding 7% under
optimal conditions and 1% under drought stress and also an increase in the frequency
of useful alleles, from 0.51 (at C0 cycle) to 0.52 (at C2 cycle).
1.8. Combining ability
Combining ability in crosses, including general combining ability (GCA) and

specific combining ability (SCA) is defined as the ability of parents to combine
amongst each other during the process of fertilization so that favourable genes or
characters are transmitted to their progenies. Evaluating combining ability by topcross method is to determine GCA and playa a very important role at the early stage
of selection when materials is too numerous. Diallel cross method is used for
evaluation of GCA and SCA of parental lines through which elite lines with high
combining ability and hybrids are selected. Besides, applying GGEBiplot for
evaluating the interaction of genotype with environment and determining
combining ability gives crucial indices: GCA, SCA effects of parents; Heterotic
groups; The best hybrids with high combining ability; The best lines .
Chapter 2. MATERIALS, ACTIVITIES AND METHODOLOGY
2.1. Materials for development of F2:3 BP populations
From 10 CIMMYT derived tropical lines showed in Table 2.1. These lines
were divided into 2 heterotic groups: Group A including elite lines P1, P2, P3, P4


7
(male parents) and line P9 (drought tolerance, the female); Group B: lines P5, P6, P7,
P8 (as the male parent) and P10 (drought tolerance, the female). Lines P9, P10 were
crossed with others in the same group into 8 F1 populations: P9×P1, P9×P2, P9×P3,
P9×P4 và P10×P5, P10×P6, P10×P7, P10×P8. Selfing F1 plants to develop 8 F2
populations, randomly selecting each cob and 100 cobs per each population and
establishing 790 families. Continue selfing these families into F2:3 generations with
the total of 790 F2:3 families represented in Table 2.2.
Table 2.1. Elite and tolerant lines

Table 2.2. Information of 8 populations and 790 F2:3 families

Notice: ǂǂPopulations developed from Bi-parents method

8
2.2. Materials
2.2.1. Materials evaluate agronomic traits and identify QTL associated with
drought tolerance of 8 F2:3 BP populations under drought stress and optimal
condition at India
8 F2:3 progenies populations were developed from each of Bi-parent cross
between drought tolerant lines P9, P10 with the elite (P1, P2, P3, P4 of Group A;
P5, P6, P7, P8 Group B) of CIMMYT (Table 2). Crossing with 2 testers CML451
(T1) and CMLO2450 (T2). Local checks: LVN10 (ĐC1), VN8960 (ĐC2), NK67
(ĐC3), C919 (ĐC4), LVN61 (ĐC5).
2.2.2. Materials for evaluating and testing hybrid combinations of 8 populations
with 2 testers (CNL451, CLO2450) in Ninh Thuan province. Including 1,605
entries: crossing 8 BP populations (790 F2:3 families) × 2 testers (CML451,
CLO2450) = 1.580 hybrid combinations; 20 hybrid combinations of 10 parental
lines × 2 testers; 5 Local checks (LVN10, VN8960, NK67, C919, LVN61).
2.2.3. Materials for researching combining ability, heterosis and drought
tolerance and grain yield of 9 elite maize lines under severe, moderate drought
stresses and optimal condition at Hyderabad, India. Including 36 diallel hybrid
combinations of 9 lines and 4 local checks (PAC745, 30V92, 900MG,
HTMH5401). These lines were original from 9 F2:3 families selected based on the
results of testing agronomic traits, crossed and developed into inbred lines and
named RA1, RA2, RA3, RA4, RA5, RA6, RA7, RA8, RA9.
Table 2.3. The list of 9 lines and 36 diallel hybrid combinations


9
2.2.4. Materials for testing maize varieties: Hybrid combinations RA2/RA8
named LVN72, RA4/RA7 named ÐH17-1.
2.3. Research contents
- Evaluate agronomic traits and identify QTL associated with drought tolerance of

8 F2:3 BP populations under drought stress and optimal condition;
- Evaluate combining ability of 8 F2:3 BP populations and select some elite lines
and promising hybrid combinations;
- Testing promising hybrid combinations.
2.4. Time and research locations
2.4.1. Research locations
- At CIMMYT, India: trials of testing 8 populations (790 families F2:3), 10 parental
lines carried out on farm under drought stress and well-watering condition; trials
of 9 lines (RA1, RA2 ... RA9) were crossed with diallel method in the field under
severe, moderate drought stresses and optimal condition.
- In Vietnam: trials for testing 8 populations x testers in the field under managed
drought stress and optimal condition; Testing promising hybrid combinations
LVN72, ĐH17-1.
2.4.2. Timelines/Milestone
During 2011- 2014 in India: testing 8 populations and parental lines in the field and
evaluating diallel crosses; During 2012-2017 in Vietnam: evaluating 8 populations
x testers and testing promising hybrids.
2.5. Research Methodology
2.5.1. Experimental designs
Under the guidance of CIMMYT and Maize Research Institute of Vietnam.
2.5.2. Methodologies of experimental evaluation in the field
- Experimental designs of evaluation in the field under drought stress and wellwatering conditions implemented by irrigation management models: wellwatering; drought; re-irrigation (as CIMMYT guidelines).
-Testing varieties according to QCVN01-56: 2011/BNN&PTNT.
- Evaluating agronomic traits under guidelines of CIMMYT and Maize Research
Institute of Vietnam.
2.5.3. Methodologies for genotyping and QTL linkage mapping
DNA extraction and linkage analysis and QTL mapping were carried out under the
guidance of CIMMYT with interval mapping method; graphs of linkage map and
QTLs expressed with MapChart software.

10
2.5.4. Processing and analysis of experimental results
Experimental data analyzed with specialized statistical softwares: Excel,
FieldBook ver.8.3, SAS ver9.0, Genstat ver12, R Studio, GGEBiplot ver4.1 and
some CIMMYT’s professional ones such as Meta 2.1 software; Analyzing and
mapping QTLs with IciMapping software, QTL CartoGrapher, MapChart.
Chapter 3
RESULTS AND DISCUSSION
Based on results of phenotyping in the field in India and Vietnam and
genotyping, 9 promising F2:3 families were selected and developed into 9 elite lines
for maize breeding adaptable to rainfed condition in Vietnam.
3.1. Evaluating agronomic traits and QTL for selecting populations and F2:3
families with drought tolerance and good combining ability
3.1.1. Evaluating agronomic traits associated with drought tolerance of 8 F2:3
populations and parental lines
* Maturity/duration
8 F2:3 populations and 10 parental lines with medium maturity; under drought
stress, these populations were in the range of 111 and 118 days, parental lines from
105 to 116 days; under optimal conditions, 8 F2:3 populations were 124 to 126 days,
lines 123 to 134 days showed table 3.1 and 3.2 showed. Anthesis date (AD) in
drought stress, populations of group A were during 68 -71 days and parental lines
from 66-76 days, group B populations from 66 to 68 days and parental line from
68 to 73 days, was not different between 2 groups.
However, anthesis silking interval (ASI) under drought stress, BP populations
(group A: 0.8 to 2.4 days, group B: 3.0 to 3.9 days) were higher than in optimal
condition (group A of 0.4-1.7, group B of 1.7-2.4 days); Besides, the variations in
ASI among 2 groups of F2:3 families were from -0.5 to 10.5 days under drought,
showed differences in the synchrony of male and female flowers of the F2:3
families.

Heritability (h2) of AD, ASI and MD (maturity days) of 8 F2:3 populations was
from low (0.1) to high (0.8). Under drought, the heritability of populations was 0.10.7/group A, 0.2-0.8/group B; Under optimal condition, group A was 0.2 - 0.7,
group B was 0.3-0.9.


11
Table 3.1. Descriptive statistics for agronomic trait recorded of group A
F2:3 populations in India

h2:Heritability; σ2g:Genotypic variance; σ2gxenv:Genotype×environment variance; ȓg:
genotypic correlation coefficients between Drought and optimal management; P9 and P 10:
Drought tolerance parent; P1 to P8 elits parent; BP: Biparental population.


12
Table 3.2. Descriptive statistics for agronomic trait recorded of group B
F2:3 populations in India

h2: Heritability; σ2g:Genotypic variance; σ2gxenv:Genotype×environment variance; ȓg:
genotypic correlation coefficients between Drought and optimal management; P9 and P 10:
Drought tolerance parent; P1 to P8 elits parent; BP: Biparental population.

* Plant and ear height and ear aspect: Plant height (PH) and ear height (EH) of 8
populations under drought stress (PH: 106-129 cm, EH: 51-67 cm) were lower than
in optimal conditions (PH: 120-140 cm, EH: 66 - 82 cm). Meanwhile, ear aspect
under drought stress was from 2.7 to 3.0 point, similar to well-watered condition
(2.6 - 2.8 point). The heritability (h2) of PH, EH and EA of 8 BP populations was
0.1 - 0.8 under drought and 0.2- 0.8 under well-watering condition.

13
* Leaf senescence: Under drought stress, leaf senescence (SEN) of all 8 F2:3 BP
populations and their parental lines was higher and higher and significant variation
at pre-flowering stages (SEN_1 from 2.2 - 2.6 point), at pollen shedding (SEN_2
from 3.8 to 5.0) and at grain-filling (SEN_3 from 5.6 to 6.2); However, under
optimal condition, the rate of leaf senescence often was to increase more slowly,
SEN_1 (0.1 - 0.2), SEN_2 (2.8 - 3.5), SEN_3 (3.7 -5. 8). The stay-green rate of
these populations under drought stress (41.5 - 46.4%) was lower than that under
full-watering (63.7 - 69.2%). The heritability (h2) of SEN_1, SEN_2, SEN_3 and
stay-green (GL) of these populations under drought stress and optimal condition
was 0.2 - 0.6 and 0.0 - 0.6, respectively.
* Lodging tolerance, rotten ears, ear tip barrenness and rate of brace roots
Stem lodging (SL) of 8 BP populations under drought stress ranged from 1.0 to
8.2%, and the parental lines were 0.4 to 17.8% was similar to that the rate under
optimal condition (SL: 0,0 - 17,1%). But ear rotten (ER) of 8 BP populations and
parent lines was generally low and small variability between drought stress and
optimal condition. Under drought stress, ER of 8 BP populations was from 1.4 to
5.1% and parent lines from 0.1 to 11.4% while, under optimal conditions, that of
the BP populations and parental lines ranged from 1.1 to 5.3% and 0.1 to 10.6%
respectively. Ear tip barrenness (TB) of BP populations under drought (scale of
2.9-3.2 score) was higher than that in optimal conditions (2.6 to 3.0 score).
Similarly, TB of parental lines (2.3 to 4.6 score) under drought condition was also
higher than that in well-watering (2.3-3.3 score). The rate of nodes for brace roots
(NBR) of 8 BP populations under drought stress in the range of 1.1 - 4.2%, parental
lines of 0.2-3.6% was higher than full watering condition (BP populations with
NBR from 0.2 - 1.8%, parent lines from 0.0 to 0.6%).
* Components of grain yield under drought stress and optimal conditions
Components of grain yield such as ear per plant (EPP), number of kernels per ear
(NK) and 1000 kernel weight (KW) of 8 F2:3 BP populations under drought stress
and optimal conditions were measured and the results showed that: under drought

stress EPP (0.6 - 0.8 ear/plant), NK (211 - 296 kernels per ear) and KW (159.7 196.9 gram) were lower than these under optimal condition with 0.8-0.9 ear/plant,
304 - 355 kernels per ear and 235.0 - 268.4 gram. The heritability (h2) of
components of grain yield including EPP, NK and KW under drought stress was
of low to medium values (0.2 - 0.7) and not different from those under optimal
condition.


14
* Productivity of 8 BP populations and parental lines under drought stress and
optimal conditions
Grain yield (GY) of 8 F2:3 populations below Table 3.11,
Table 3.11. Descriptive statistics for grain yield recorded
of F2:3 populations in India

h2: Heritability; σ2g:Genotypic variance; σ2gxenv:Genotype×environment variance; ȓg: genotypic correlation
coefficients between Drought and optimal management; P9 and P 10: Drought tolerance parent; P1 to
P8 elits parent; BP: Biparental population.


15
Under drought stress (0.74 - 1.37 tons.ha-1) was lower than that under optimal
condition (1.21 - 2.51 tons.ha-1). The grain yield of parental lines under drought
stress ranged from 0.17 to 1.31 tons.ha-1, lower than the productivity in wellwatering condition (0.56-2.38 tons.ha-1). It was found that grain yield under
drought stress decreased sharply if comparing with optimal conditions by 22.0 48.6% for 8 populations, 8.4-78% for parent lines. But the reduction in grain yield
of drought tolerant lines P9, P10 used as the female was smaller than lines P1 to
P8 as the male did). Meanwhile, the heritability (h2) of populations was from
medium to high and naturally it was lower under stresses than under optimal
condition. The heritability of group A was 0.21-0.55 and 0.75 - 0.85 while group
B was 0.55-0.69 and 0.51-0.89 under drought and full irrigation, respectively. In
brief, 8 F2:3 BP populations (BP1, BP2 ... to BP8), developed from 10 inbred lines

including 8 male ones (P1, P2, P3, P4, P5, P6, P7, P8) and 2 female ones (P9, P10),
were significantly different in traits. Under drought stress, female lines (P9, P10)
were better at drought tolerance than the elite ones P1, P2, P3, P4, P5, P6, P7, P8.
The variation of F2:3 families in each population expressing the segregation,
probably over parents in the positive or negative trends in traits of AD, ASI, PH,
EH, SEN_2, SEN_3, TB, green leaves, EPP, NK, KW and GY showed if the
segregation of allels was favorable or not, through which useful allels and good F2:3
families were selected for maize breeding.
3.1.2. Correlation coefficient of phenotype and genotype on agronomic traits
and grain yield of 8 BP populations under drought stress and optimal
conditions
Correlation coefficient of phenotype of agronomic traits (AD, ASI, EA, PH, EH,
TB, NBR, NK, KW) and grain yield - GY under drought stress and optimal
conditions was positive while traits of MD, SEN_2 and SEN_3, SL, ER were of
negative correlation with the range of - 0.5 and -0.12. Leaf senescence at the preflowering (SEN_1), ER and EP were not close correlation. But the correlation of
traits (AD, ASI, PH, NBR, NK, GL and GY) under drought stress and optimal
condition was positive, which showed that these traits were associated with
genome regions controlling their attributes and suggest identifying common QTLs;
Except for traits of SEN_2, SEN_3 with negative correlation (from -0,64 to -0,29)
under drought stress and optimal condition. The reason is that the genetic variation


16
of populations developed from drought tolerant lines with the elite ones were
dominant or additive on these agronomic traits. Therefore, traits of AD, ASI, PH,
SEN_2 and GY may be used for researching drought tolerance of 8 populations.
3.1.3. Mapping genome regions regulating drought tolerance of 8 F2:3
populations
There were 871 SNPs identified for mapping 7 BP populations (BP1, BP2, BP3,
BP4, BP5, BP6, BP7) showed on Figure 3.4. Meanwhile BP8 was of no SNP

polymorphism, so not being mapped. QTLCartographer v2.5 software was used
for QTL mapping. Genome regions controlling traits of GY, ASI and SEN on 10
chromosomes of 7 populations (BP1, BP2, BP3, BP4, BP5, BP6, BP7) were
located (bin). 2 QTL clusters were found: the first was on chromosome 1 (bin 1.051.07) with QTLs associated with ASI due to additive effects and grain yield under
drought stress by additive, dominant and over-dominant effects, and chromosome
7 (bin 7.01-7.03), including 3 QTLs related to ASI under drought stress and the
another controlling grain yield and ASI. the second was on chromosome 8 (bin
8.01-8.03), QTLs regulating traits of ASI and SEN under drought stress.

Figure 3.4. Maped QTL for GY, ASI and SEN on 7 population F2:3 conected


17
The other was on chromosome 8 (bin 8.01-8.03) with QTLs regulating traits of ASI
and SEN under drought stress. In addition to these clusters of QTLs, other smaller
clusters, on chromosome 4 (bin 4,06-4,08) and on chromosome 6 (6.05-6.07 bin)
finding 2 QTLs for ASI, SEN and GY under drought stress; these genome regions
were identified to be incomplete dominant, except for few QTL with clear
expression of additive effects. It was showed that useful alleles were derived from
drought tolerant lines (P9, P10) and the elite ones (P1, P2 ..., P8). However, in
maize breeding, it would be not only selecting good materials but also evaluating
their combining ability for breeding maize for production.
3.2. Evaluating for combining ability of 8 F2:3 populations, Selecting elite lines
and promising hybrid combinations
3.2.1. Early testing analysis for combining ability in of 8 F2:3 populations
Trials of evaluation of test-crosses by diallel of 8 populations (790 F2:3 families)
and 10 parental lines with 2 testers (CML451 and CLO2450) under managed
drought stress and optimal condition were implemented in Spring 2014 at Nha Ho,
Ninh Thuan showed on Table 3.28.
a, Grain yield of F1 hybrids in diallel crosses under drought stress and optimal

condition at Spring 2014 in Ninh Thuan
Results showed that reduction in grain yield of the F1 hybrids of [BP populations
× testers] and [parental lines × testers] under drought stress was by 27.23 - 54.16%
for [F2:3 × testers], 16.22 - 100,00% for [parental lines × testers] of group A and
5.88-83.30% for [parental lines × testers] of group B, if comparing with optimal
condition. The variation in grain yield occurred in each segregating population BP1
to BP8 under both drought stress and optimal conditions, moreover some [F2:3
families× testers] were of higher or lower grain yield than that of [parental line ×
testers], which approved that some F2:3 families were inherited from their parental
lines and their combining ability was improved through diallel cross with grain
yield over parents.
The productivity of [F2:3 families × testers] under drought was the highest with 6.89
tons.ha-1 and in optimal condition with 8.29 tons.ha-1, higher than the yield of 5
checks of LVN10 (C1), VN8960 (C2), NK67 (C3), C919 (C4), LVN61 (C5) with
3.15-5.00 tons.ha-1 under drought stress and 4.79 - 7.36 tons.ha-1 in optimal
conditions, reduction by 8.02-35.62%, respectively. The results showed that,
among 790 F2:3 families of 8 BP populations, through selection and development,
there are some useful ones with the genetic improvement of drought tolerance,


18
higher in grain yield than parents, especially over or equivalent to 5 local checks,
selected as materials for maize breeding toward rainfed condition.
b. Evaluating combining ability for grain yield of 8 F2:3 populations and parental
lines under drought stress and optimal condition.
General combining ability (GCA) of [BP populations × testers] under drought
stress and well-watering was ranged -2.23 to 2.64 and -2.35 to 2.94 for group A;
2,38 to 2.64 and -1,70 to 2.94 for group B, respectively whereas the GCA of the
parental lines ranged from -1.64 to 2.37 under drought stress, from -4.50 to 1.52 in
optimal conditions;

Under stress, SCA of [populations in group A × testers] and [populations in
group B × testers] was -1.72 to 2.94, -2.28 to 2.30 respectively. If being wellwatered, it was -2,38 to 2.30 for [populations in group A × testers], -2.17 to 2.18
for [populations in group B × testers] while the SCA of [parental lines × testers]
was from -1.33 to 1.35 under drought stress, from -1.06 to 1.07 in optimal
conditions.
It was approved that SCA under drought stress was higher than that in optimal
condition and all F2:3 families of each BP population were of segregation.
Moreover, under both these conditions, SCA of [8 populations with testers] were
all higher than that of [lines with testers].
3.2.2. Selecting elite lines and promising hybrid combinations
Based on the results of analyzing drought tolerance, identifying QTLs associated
with drought tolerance and combining ability of 790 F2:3 families 8 BP populations
under drought stress and optimal condition, F2:3 families were selected with not
only combining ability, drought tolerance but also high grain yield for drought
tolerant maize breeding. 54 F2:3 families with high GCA and SCA were selected
and under drought stress and optimal condition their grain yield was 3.91 - 5.92
tons.ha-1 and 5.76-8.63 tons.ha-1 respectively, higher or equivalent to 5 local checks
(3.15-5.00 tons.ha-1, 4.79-7.36 tons.ha-1) and [parental lines x testers] (0,0-5.08
tons.ha-1; 0.03 – 6.05 tons.ha-1).
Based on drought tolerance, QTL analysis and early generation testing for
combining ability of 8 F2:3 populations, there were selected 9 F2:3 families, BP1_46,
BP1_74, BP2_109, BP3_41, BP4_40, BP5_85, BP6_72, BP7_10, BP8_21, with
good general combining ability and specific combining ability under both drought


19
stress and optimal conditions and then developed into inbred lines named RA1,
RA2, RA3, RA4, RA5, RA6, RA7, RA8, RA9, which were improved and
crossed in diallel for evaluation and selection of elite hybrid combinations for
production.

Table 3.24. Characteristic 9 lines selected, development and maintained

3.2.3. Grain yield of lines and hybrid combinations under experimental
conditions
Inbred lines named RA1, RA2, RA34, RA4, RA5, RA6, RA7, RA8, RA9 were
crossed to develop 36 diallel-crosses. These test-crosses and lines were tested with
4 commercial varieties (PAC745, 30V92, HTMH5401 and 900MG) under severe,
medium drought stresses and optimal conditions in 2014 in India. The average
yield of 9 lines reached 3.9 tons.ha-1 under optimal conditions, 0.7 tons.ha-1 under
severe drought stress (reduced by 83%) and 1.4 tons.ha-1 in medium stress
(decreased by 63%.) It was found that F1 hybrids showed better drought tolerance
than that of parental lines because it was only reduced 50-93% under severe
drought stress, 10 - 53% under medium drought stress. The average yield of hybrid
combinations under well-irrigating condition was 7.2 tons.ha-1, of which some of
them such as RA1/RA8, RA6/RA9, RA6/RA8 were of 8.7 - 8.8 tons.ha-1,
equivalent to the check 900MG (8.0 tons.ha-1).
Under severe drought stress, the average yield was 2.1 tons.ha-1, but hybrids
RA4/RA7 and RA7/RA9 was 3.5 tons.ha-1, equal to the productivity of the check


20
30V92 (3.1 tons.ha-1). Under medium drought stress, hybrid combination
RA2/RA8 6.0 tons.ha-1, equivalent to the check HMH5401 (5.7 tons.ha-1).
3.2.4. Combining ability of 9 new lines under experimental conditions
Under all conditions (moderate severe drought stress and well-watering), it was
showed that: Under well-watered conditions: for line RA8 (0.9 tons.ha-1), GCA
effects for grain yield was significant and reached 4.7 tons.ha-1, followed by lines
RA9 (GCA = 0.6) and RA6 (GCA = 0.5). Under severe drought stress: line RA7
had the highest GCA effect (GCA = 0.6), then line RA8 (GCA = 0.5) and line RA9
(GCA = 0.4). Under moderate drought stress: GCA effects of line RA8 (GCA =

0.5) and RA9 (GCA = 0.3), next to lines RA2, RA4 with GCA = 0.2 were
significant at P <0.01. Thus, these lines may make the grain yield of hybrid
combinations higher than the average productivity of three conditions.
On specific combining ability (SCA), it was showed: Under optimal
conditions: hybrid RA3/RA5 (SCA = 0.9 with 6.6 tons.ha-1), RA3/RA7 (SCA =
0.8 with 6.9 tons.ha-1 and RA2/RA8 (SCA = 0.6 and 8.1 tons.ha-1) were equivalent
to the check 900MG (8.0 tons.ha-1); Under severe drought stress: hybrid RA3/RA5
reached the highest SCA as 0.9 and the highest yield as 3.7 tons.ha-1, RA6/RA7
was 3.3 tons.ha-1, RA2/RA9 had SCA of 0.8 and 3.0 tons.ha-1.
The grain yield of these hybrids were equivalent to or over the check 30V92
(3.1 tons.ha-1); Under moderate drought stress: Hybrid RA5/RA6 was the highest
in SCA (0.9) with 5.2 tons.ha-1. Hybrids RA2/RA8, RA6/RA9 and RA2/RA7 also
had relatively high SCA effects with 6.0 tons.ha-1, 5.7 tons.ha-1 and 5.4 tons.ha-1
respectively, equivalent to the check HMH5401 (5.7 tons.ha-1). Thus, it may be to
select hybrids RA3/RA5, RA4/RA7 for well-watered conditions and severe
drought stress; hybrid RA5/RA6, RA2/RA8 for moderate drought stress and
optimal conditions.
3.2.5. Analyzing interaction between genotype and environment
With applying GGEbiplot for analysis of results under 3 conditions, it was showed
that: In well-watered conditions, line RA8 had the highest GCA effect, followed
by lines RA9 and RA6, finally lines RA5, RA2 and RA3; Under severe drought
stress: lines RA4, RA6 and RA9 were formed as the first group, the next consisted


21
of lines RA1, RA5, RA8 and RA7; And line RA2 was considered the third group;

Under intermediate drought stress: lines RA3, RA2, RA1 and R8 was of location
at the top of the polygon of a group, which showed that these lines can combined
well together. Thus, lines RA2 and RA8 were identified as the highest heterosis

22
(RA2/RA8 with grain yield of 6.0 tons.ha-1 and SCA = 0.6). RA4/RA7 was the
highest in grain yield under drought (3.5 tons.ha-1) and its average under 3
experimental conditions was 5 tons.ha-1 and reduction in yield under moderate
drought (34%) and severe drought (50%) was the lowest among hybrids. Under
research cooperation program by CIMMYT; Vietnam is a member and so it was
allowed for us to use maize materials and germplasm of this program for research
since 2012 for development into elite inbred lines for domestic maize breeding.
Based on the results of researches in India and in Vietnam, selection of lines and
development of hybrids were implemented in Vietnam. During the development
of inbred lines and hybrids in Vietnam, hybrid combinations called RA4/RA7 (as
ĐH17-1), RA2/RA8 (as LVN72) always showed high grain yield, in which hybrid
was named.
3.3. Results of testing maize hybrid LVN72, ĐH17-1 in Vietnam
3.3.1. Results of testing promising hybrid combination LVN72
Hybrid combination LVN72 has been tested in the testing network of National
Center for Plant Testing in the Autumn-Winter 2016 in the Red River Delta, the
Northern Mountainous Regions and the North Central Region. The results showed
that days to anthesis of LVN72 was from 56 to 79 days; to silking of 58 to 75 days;
to physiological maturity of 111 to 117 days, equivalent to the local check DK9901
(108 - 117 days); Drought and cold tolerance was as good as DK9901 (score of 1
to 3).
Grain yield of LVN72 was 6.6 to 7.8% higher than that of the local check in
the Red River Delta and 7.3-20.6% in the Northern Mountainous Regions.
Especially, in Son La province, at Spring- Summer crop season, LVN72 reached
11.2 tons.ha-1, over the check by 11.2% at the significant confidence level.
3.3.2. Results of testing promising hybrid combination ĐH17-1
In the Spring crop season 2017, at Dan Phuong, Hanoi, days to anthesis was

74 days, silking of 75 days; to physiological maturity of 119 days, equivalent to the
check DK9901 and NK67 and longer than LVN99. Grain yield of ĐH17-1 was as
high as checks DK9901 (ĐC1), NK67 (ĐC2), LVN99 (ĐC3). ĐH17-1 with 10.2
tons.ha-1, thus this hybrid combination is promising and necessary to be tested more
for development in production.


23
CONCLUSION AND SUGGESTION
I. Conclusion
1. Positive genetic correlations were observed in some traits such as time from
sowing to anthesis date (AD), anthesis-silking interval (ASI), plant height (PH),
brace roots (NBR), number of kernel, stay green and grain yield (GY); negative
correlation in leaf senescence (SEN) for drought tolerance. Eight new population
progenies of the F2:3 generation were identified for better than parents for all
traits. Having identified grain yield of 8 F2:3 populations (0,74 - 1,37 tons.ha-1),
22 - 48.6% lower than that under optimal conditions (from 1.21 to 2.51 tons.ha1

) and less than of the parental lines (8.4- 78%), which showed that they were

better in drought tolerance than their parents did. The variation on traits of AD,
ASI, PH, EH, SEN, KW, GL, TB and GY in F2:3 populations took place under
drought stress and optimal condition, the values on grain yield of families in each
population were positive (higher than that of parents) or negative (lower than
parents’) under drought stress and well-watering, based on for selection of F2:3
families for maize breeding.
2. There were 27 QTLs identified associated with drought tolerance on 3 traits of
ASI, SEN and GY, of which 11 QTL_GY, 6 QT_ASI and 10 QTL_SEN. The
cluster identified on chromosome 1 (bin 1.05-1.07) with QTLs associated with
drought tolerance on ASI and GY may be applied into drought tolerant maize

breeding with molecular markers. Other QTLs found on chromosome 4 (bin
4.03-4.05) related to GY and chromosome 7 (bin 7.01-7.03) linked to ASI can be
used to identify the combination of dorught tolerant materials with elite
germplasm. In addition, the cluster located on chromosome 8 (bin 8.02-8.03)
including QTLs for ASI and SEN and chromosome 6 (bin 6.05-6.07) with QTLs
controlling GY and SEN are probably developed for researching maize tolerant
to drought.
3. Through diallel crossing, it was showed that GCA and SCA for grain yield of 8
F2:3 populations, 10 parental lines and 2 testers (CML451, CLO2450) under
drought stress were higher than those in optimal conditions, which approved that
additive effects play a more important role in grain yield while dominant ones
are significant. Moreover, SCA, GCA and productivity of [F2:3 populations x
testers] were of bigger variation more than [lines x testers]. Drought tolerance of