|
African Crop Science Journal
African Crop Science Society
ISSN: 1021-9730 EISSN: 2072-6589
Vol. 8, Num. 3, 2000, pp. 223-232
|
African Crop Science Journal, Vol. 8. No. 3, pp. 223-232
African Crop Science Journal, Vol. 8. No. 3, pp. 223-232
GENETIC ANALYSES OF ADAPTATION OF BARLEY TO LOW AND HIGH
INPUT ENVIRONMENTS
R. MADAMBA Laboratory of Plant Breeding, Wageningen Agricultural University, The Netherlands Present mailing address: Department of Research and Specialists Services, Crop Breeding Institute, Box CY 550, Causeway, Harare, Zimbabwe
(Received 17 May, 1999; accepted 2 May, 2000)
Code Number: CS00024
INTRODUCTION
Genetic adaptation implies the shaping of a population or a
species gene pool in response to environmental challenges (Perez-de-la-Vega
and Tigerstedt, 1996). A crops ability to exploit its environment depends on
many adaptive features that are controlled by multiple genes, interacting among
themselves and with the environment in complex ways (Hawtin et al., 1997).
The genotype by environment interaction (GxE) observed by plant breeders indicates
differential responses of the varieties being tested to different environmental
conditions and is a major complication in plant breeding (Ceccarelli and Hammer,
1996). GxE interactions reduce the correlation between the phenotype and the
genotype.
Ceccarelli et al. (1996) suggested that if the genotype
by environment (GxE) interaction is of the crossover type, genotypes developed
in favourable environments do not perform well under harsh environments. This
suggests that the genes for yield expressed in low and high input conditions
are different. As a result, breeding procedures conducted under high input and
uniform conditions might favour selection of cultivars adapted to intensive
management and might eliminate individuals adapted to low input conditions (Ceccarelli,
1997).This could be the reason why research on crop improvement has not had
as much impact in the smallholder farms compared to commercial farms. The idea
of breeding for specific environments rather than broad adaptation is supported
by Falconer (1990) and Kearsey and Pooni (1996).
Braun et al. (1997) however, suggested that it
is possible to breed for wide adaptation provided that the genetic base is broad
enough. The use of a shuttle programme (shuttling populations between localities
with widely different environments) enables simulta-neous selection for both
high and low yielding environments.
It was against this background that we initiated an experiment
to obtain more insight into the genetic background of adaptation, both to specific
and broad conditions at the molecular level. Knowledge of quantitative trait
loci by enviro-nment interaction (QTL x E) is essential to be able to use the
correct markers for marker assisted selection. The objective of the present
study was to observe whether the same QTLs are involved in low, intermediate
and high input environments (QTL by environment interaction) and whether the
information on QTL by environment interaction can give some insights on breeding
for specific and broad adaptation
MATERIALS AND METHODS
Ninety-four barley recombinant inbred lines (RILs), being progenies
of reciprocal crosses of Apex and Prisma and the two parents, were used in this
study. Apex contains the recessive mlo-mutant allele and is powdery mildew resistant
and Prisma is susceptible to powdery mildew. Prisma contains the denso mutant
allele that gives it a short stature.
Trials were conducted in a field with sandy soil at the Wageningen
Agricultural University farm. It was grown in three different input environments
namely; 1) 90 kg ha-1 total nitrogen + pesticides, representing the high input
environment (NP); 2) 0 kg ha-1 applied nitrogen + pesticides, representing
the intermediate input environment (P); and 3) 0 kg ha-1 applied nitrogen and
no pesticides, representing the low input environment (O). The results of soil
analysis indicated 26 mg (P2O5) l-1 of phosphorus, 13 mg (K2O)100 g-1 of potassium,
166 mg (MgO) kg-1 of magnesium and 0.31 mg (B) kg-1 of boron and these were
deemed adequate for barley growth. The results for nitrogen were 35 kg N ha-1.
The rates of nitrogen applied in the high input environment were based on the
results from the soil analysis.
The design used was a randomised incomplete block design (a-design)
in a split plot arrangement, with the fertiliser-pesticide levels as main plots
and was replicated three times. The gross and net plots were 9.0 m2 . The spacing
between sub-plots was 30 cm, but 3 m for the main plots.
Seventy-five kg N ha-1, in the form of Kas (Kalli ammon salpeter)
at a rate of 275 kg ha-1, were applied on the 15 May, on plots that were to
receive nitrogen to bring the total nitrogen (residual nitrogen was 35 kg ha-1)
in the fertiliser-pesticide treatment to 90 kg N ha-1. The application rate
of 75 kg N ha-1 instead of 55 kgha-1 was based on the efficiency of the machine.
The Kas fertiliser contains 27% N of which 13.5% is in nitrate form for a long
term effect and 13.5% in form of ammonium for immediate use. The trial was sprayed
against powdery mildew with amistan (azoxystropin) and opus team (epoxiconazool/fen
propimorf) at rates of one litre and one and half litres per hectare, respectively.
The trial was sprayed against aphids with pirimor (pirimicarb) at a rate of
0.4 kg ha-1. The experiment received 3 l ha-1 of verigal D (bifenox/mecoprop-P),
a herbicide against broad leafed weeds. The trial was planted on 31 March and
1 April 1998, at seeding rate of 90 kg ha-1 and was harvested on August 11 and
12, 1998.
The records taken included; grain yield in grams per plot,
number of kernels per ear, weight of thousand kernels, number of ears per square
metre, days to 50% flowering (the number of days from planting to the time when
50% of the main tilllers in a plot have exposed awns), days to 95% maturity
(the number of days from planting, required by 95% of the plants to reach physiological
maturity), lodging score and plant height. Lodging score was obtained as a fraction
of the area lodged x lodging severity, where lodging severity is scored from
0 to 2; 0 represents no lodging, 1 plants bent but not touching the ground,
and 2 plants are completely laying prostate on the ground.
A general analyses to test for outliers, equality of residual
variance, genotype by environment interaction and generation of means was carried
out on the phenotypic data using the statistical package SAS copyright 1989-1995
of the SAS institute. ANOVA was conducted for each environment and for combined
environments. ANOVA and least squares means (means adjusted for the effect of
incomplete blocks) for both the individual and combined analyses (for data with
equal variances) were generated using the PROC GLM procedure. The PROC GLM procedure
with the weighted command was used for the data with unequal variances. Bartletts
and Levenes tests (Gomez and Gomez, 1984; Snedecor and Cochran, 1989), were
used to test for homogeneity of variance.
A genetic linkage map, which was constructed from the barley
recombinant inbred lines (progenies of Apex and Prisma) at the Laboratory of
Plant Breeding (Dourleijn, pers. comm.) was used for QTL analyses using the
computer program MapQTL version 3.0, copyright 1997 of CPRO-DLO, Wageningen.
The linkage map was constructed from amplified fragment length polymorphism
(AFLP) markers. The AFLP used was constructed using 190 AFLP markers from a
population of 94 recombinant inbred lines.
RESULTS
The residuals of the data for all the traits in the three different
environments showed no deviation from a normal distribution. Variances for number
of days to 50% flowering (DTF), plant height (PH) and grain yield over the three
different environments were not significantly different. The variances for the
data on number of days to 95% physiological maturity (DMAT), weight of a thousand
kernels (TKW), number of kernels per ear (KE) and number of ears per square
metre (EARM2) differed significantly (P<0.05). Both the logarithm and square
root transformations failed to make the variances equal.
The individual analyses of variances (data not shown) showed
highly significant (P<0.05) line effects for all the traits measured in all
the three environments. ANOVA also indicated significant blocking effects for
all the traits in each of the individual environments. The combined analyses
of variance of all the treatments showed significant genotype by environment
interactions effects for all the traits measured except DTF.
Correlation coefficients for the traits measured in each
of the three different environments. Grain yield (KGHA) was positively correlated
with DTF, TKW, and EARM2 and negatively correlated with lodging score (LOG)
in the NP environment (Table 1). Grain yield (KGHA) was positively correlated
with DTF, DMAT and EARM2 in the P environment and positively correlated with
TKW and EARM2 in the O environment. High negative correlations were observed
between KE and EARM2 and TKW and EARM2 in all the three environments. The correlations
between KE and TKW were positive. High positive correlations were also found
between PH and KE and DMAT and DTF in all the three environments and between
LOG and PH in the NP environment.
TABLE1. Correlation coefficients/Prob|R| under HO: RHO = 0/N
= 100 in different environments |
|
|
KGHA1 |
KE |
EARM2 |
TKW |
DTF |
PH |
DMAT |
LOG |
KE |
NP |
0.13107 |
|
|
|
|
|
|
|
|
|
0.19372 |
|
|
|
|
|
|
|
|
P |
-0.05791 |
|
|
|
|
|
|
|
|
|
0.5671 |
|
|
|
|
|
|
|
|
O |
-0.03464 |
|
|
|
|
|
|
|
|
|
0.7323 |
|
|
|
|
|
|
|
EARM2 |
NP |
0.27046 |
-0.78023 |
|
|
|
|
|
|
|
|
0.006.5 |
0.0001 |
|
|
|
|
|
|
|
P |
0.46036 |
-0.80932 |
|
|
|
|
|
|
|
|
0.0001 |
0.0001 |
|
|
|
|
|
|
|
O |
0.48609 |
-0.77451 |
|
|
|
|
|
|
|
|
0.0001 |
0.0001 |
|
|
|
|
|
|
TKW |
NP |
0.22092 |
0.41385 |
-0.68666 |
|
|
|
|
|
|
|
0.0272 |
0.0001 |
0.0001 |
|
|
|
|
|
|
P |
0.08524 |
0.34988 |
-0.58413 |
|
|
|
|
|
|
|
0.3991 |
0.0004 |
0.0001 |
|
|
|
|
|
|
O |
0.22625 |
0.35604 |
-0.53183 |
|
|
|
|
|
|
|
0.0236 |
0.0003 |
0.0001 |
|
|
|
|
|
DTF |
NP |
0.205588 |
0.00782 |
0.11103 |
-0.0243 |
|
|
|
|
|
|
0.0399 |
0.9384 |
0.2714 |
0.8104 |
|
|
|
|
|
P |
0.23144 |
-0.18307 |
0.30786 |
-0.21942 |
|
|
|
|
|
|
0.0205 |
0.0683 |
0.0018 |
0.0283 |
|
|
|
|
|
O |
-0.07962 |
-0.14107 |
0.21894 |
-0.36606 |
|
|
|
|
|
|
0.431 |
0.1615 |
0.0286 |
0.0002 |
|
|
|
|
PH |
NP |
-0.1764 |
0.45419 |
-0.47942 |
0.21842 |
-0.57203 |
|
|
|
|
|
0.0791 |
0.0001 |
0.0001 |
0.029 |
0.0001 |
|
|
|
|
P |
-0.11484 |
0.53513 |
-0.50489 |
0.33201 |
-0.67103 |
|
|
|
|
|
0.2552 |
0.0001 |
0.0001 |
0.0007 |
0.0001 |
|
|
|
|
O |
0.16239 |
0.46519 |
-0.415 |
0.53334 |
-0.66602 |
|
|
|
|
|
0.1065 |
0.0001 |
0.0001 |
0.0001 |
0.0001 |
|
|
|
DMAT |
NP |
0.06825 |
0.00729 |
-0.07444 |
0.17008 |
0.49967 |
-0.14198 |
|
|
|
|
0.4999 |
0.9426 |
0.4617 |
0.0907 |
0.0001 |
0.1588 |
|
|
|
P |
0.40106 |
-0.43384 |
0.56198 |
-0.2463 |
0.6887 |
-0.64747 |
|
|
|
|
0.0001 |
0.0001 |
0.000 |
0.0135 |
0.0001 |
0.0001 |
|
|
|
O |
0.05444 |
-0.31985 |
0.38042 |
-0.31208 |
0.63832 |
-0.54392 |
|
|
|
0.5906 |
0.0012 |
0.0001 |
0.0016 |
0.0001 |
0.0001 |
|
|
|
LOG |
NP |
-0.3058 |
0.33718 |
-0.38166 |
0.06499 |
-0.46196 |
0.75323 |
-0.17395 |
|
|
|
0.002 |
0.0006 |
0.0001 |
0.5206 |
0.0001 |
0.0001 |
0.0835 |
|
1KGHA = grain yield, KE = number of kernels per
ear, 2EARM = Number of ears per square metre
TKW = weight of a thousand kernels, DTF = days to 50% flowering, PH = plant
height
DMAT = days to physiological maturity, LOG = lodging score. Numbers in italics
are the probability values |
QTL analyses. Table 2 shows the results of QTL analysis
for the traits measured in the three different environments. Five QTLs affecting
days to 50% flowering (DTF) were detected on chromosomes 1, 2, 3 (2 QTLS) and
5, in the NP environment. Six were detected on chromosomes 1, 2, 3 (2 QTLs),
4 and 5, in the P environment and also six in the O environment on chromosomes
1, 2, 3, 4 (2 QTLs) and 5. Chromosome 3 at a position of about 126cM appeared
to have a major effect on days to 50% flowering in all the three environments.
The phenotypic variation explained by the QTL in this region was 56.9% for NP
and 50.5% and 58.6% for the P and O environments, respectively. The additive
effect of this QTL in the NP environment was 1.7 days, 1.5 days for the P environment
and 1.8 days for the O environment.
TABLE 2. Results of Multiple QTL Mapping (MQM) for different
agronomic traits |
3Trait |
1Inputs |
Chromosome |
QTL (cM) position |
2Additive effects |
%Variation explained |
Lod score |
DTF |
NP |
1 |
30 |
0.9 |
17.5 |
8.85 |
2 |
120 |
0.5 |
4.8 |
3.78 |
3 |
126 |
1.7 |
56.9 |
22.59 |
3 |
65 |
-0.5 |
4.0 |
2.82 |
5 |
104 |
-0.6 |
6.6 |
4.51 |
P |
1 |
30 |
0.9 |
17.1 |
7.33 |
2 |
120 |
0.5 |
4.8 |
3.16 |
3 |
126 |
1.5 |
50.5 |
19.40 |
3 |
65 |
-0.4 |
3.3 |
2.20 |
4 |
6 |
-0.4 |
3.1 |
2.05 |
5 |
104 |
-0.9 |
5.6 |
3.27 |
O |
1 |
22 |
1.0 |
16.1 |
8.80 |
2 |
119 |
0.7 |
6.8 |
4.75 |
3 |
126 |
1.8 |
58.6 |
21.58 |
4 |
8 |
-0.5 |
5.5 |
3.11 |
4 |
79 |
-0.6 |
5.7 |
3.77 |
5 |
100 |
-0.4 |
3.2 |
2.34 |
DMAT |
NP |
1 |
26 |
0.6 |
19.2 |
7.35 |
3 |
126 |
0.7 |
17.9 |
5.10 |
3 |
145 |
-0.4 |
8.5 |
3.27 |
4 |
67 |
-0.4 |
7.8 |
3.31 |
7 |
75 |
-0.5 |
11.1 |
4.60 |
P |
1 |
30 |
0.9 |
12.7 |
3.86 |
3 |
126 |
1.8 |
51.5 |
10.96 |
7 |
75 |
-0.6 |
5.6 |
2.10 |
O |
1 |
35 |
0.8 |
13.7 |
3.04 |
3 |
126 |
1.5 |
43.9 |
13.39 |
7 |
86 |
-0.6 |
7.1 |
3.12 |
PH |
NP |
1 |
22 |
2.8 |
7.1 |
6.14 |
1 |
81 |
2.6 |
5.9 |
3.80 |
3 |
126 |
-9.2 |
73.3 |
29.88 |
4 |
129 |
2.7 |
6.3 |
3.62 |
5 |
105 |
2.4 |
5.0 |
2.97 |
P |
1 |
23 |
1.3 |
1.3 |
2.16 |
1 |
126 |
1.9 |
2.8 |
3.56 |
2 |
68 |
1.3 |
1.3 |
2.17 |
2 |
120 |
-2.1 |
3.2 |
4.74 |
3 |
126 |
-9.6 |
75.1 |
40.87 |
4 |
129 |
2.0 |
3.1 |
2.59 |
5 |
104 |
2.1 |
3.4 |
5.25 |
O |
1 |
22 |
1.8 |
2.3 |
3.23 |
1 |
85 |
1.9 |
2.8 |
3.95 |
1 |
126 |
2.1 |
3.3 |
4.31 |
2 |
68 |
1.6 |
1.9 |
3.20 |
3 |
126 |
-10.1 |
76.3 |
41.11 |
4 |
129 |
2.1 |
3.1 |
3.44 |
5 |
103 |
2.8 |
4.9 |
6.96 |
EARM2 |
NP |
1 |
40 |
-52.0 |
33.5 |
3.46 |
2 |
125 |
-39.6 |
18.9 |
4.49 |
3 |
126 |
36.6 |
17.5 |
4.41 |
4 |
141 |
-41.3 |
21.3 |
4.85 |
6 |
93 |
-36.6 |
17.3 |
2.80 |
P |
2 |
73 |
-24.4 |
19.0 |
4.92 |
3 |
126 |
26.9 |
23.6 |
7.21 |
4 |
129 |
-25.6 |
20.1 |
4.33 |
7 |
30 |
18.6 |
8.3 |
3.45 |
7 |
76 |
13.8 |
6.2 |
2.06 |
O |
1 |
40 |
25.7 |
14.9 |
3.66 |
2 |
130 |
-15.1 |
7.4 |
3.28 |
3 |
126 |
24.3 |
20.9 |
6.99 |
4 |
134 |
-24.4 |
18.9 |
5.12 |
6 |
67 |
-15.6 |
8.5 |
2.10 |
7 |
26 |
16.7 |
7.0 |
3.28 |
7 |
76 |
17.7 |
11.2 |
3.31 |
KE |
NP |
1 |
35 |
0.7 |
19.7 |
3.28 |
2 |
120 |
0.7 |
18.5 |
6.77 |
3 |
126 |
-0.8 |
12.3 |
4.23 |
4 |
141 |
0.4 |
8.1 |
2.31 |
7 |
30 |
-0.5 |
10.9 |
3.98 |
P |
2 |
130 |
0.6 |
12.0 |
5.16 |
2 |
60 |
0.4 |
10.9 |
3.39 |
3 |
60 |
-0.5 |
9.3 |
2.99 |
3 |
126 |
-0.5 |
8.6 |
3.27 |
4 |
124 |
0.7 |
17.9 |
3.84 |
7 |
20 |
-0.5 |
7.5 |
3.40 |
O |
2 |
130 |
0.7 |
20.0 |
7.47 |
2 |
65 |
0.4 |
7.5 |
3.17 |
3 |
60 |
-0.5 |
11.5 |
3.27 |
3 |
126 |
-0.4 |
8.7 |
3.30 |
7 |
30 |
-0.4 |
8.4 |
3.39 |
TKW |
NP |
1 |
85 |
1.0 |
11.0 |
4.27 |
2 |
120 |
0.9 |
8.4 |
3.63 |
6 |
81 |
0.9 |
9.9 |
3.15 |
7 |
22 |
-1.0 |
7.7 |
3.43 |
7 |
76 |
-0.9 |
10.2 |
4.21 |
P |
1 |
85 |
0.9 |
8.9 |
3.75 |
2 |
65 |
0.8 |
6.8 |
3.12 |
3 |
126 |
-1.0 |
13.2 |
5.13 |
4 |
79 |
-0.9 |
9.0 |
3.11 |
6 |
41 |
1.0 |
11.6 |
5.09 |
O |
1 |
85 |
0.9 |
9.4 |
4.75 |
2 |
68 |
1.0 |
12.1 |
5.96 |
3 |
126 |
-1.3 |
21.3 |
9.47 |
4 |
75 |
-0.7 |
5.8 |
3.38 |
4 |
152 |
-0.8 |
5.7 |
3.29 |
6 |
44 |
1.0 |
11.3 |
6.00 |
7 |
76 |
-0.6 |
5.0 |
2.12 |
KG/HA |
NP |
3 |
126 |
273.1 |
38.0 |
13.12 |
3 |
170 |
110.2 |
6.4 |
2.68 |
4 |
134 |
-213.7 |
22.2 |
7.43 |
7 |
0.0 |
-134.0 |
8.7 |
4.45 |
7 |
101 |
-122.6 |
7.8 |
3.80 |
P |
1 |
35 |
202.1 |
45.9 |
9.43 |
3 |
126 |
124.3 |
18.0 |
6.19 |
4 |
136 |
-106.5 |
12.8 |
4.63 |
6 |
41 |
83.9 |
8.5 |
3.55 |
O |
1 |
30 |
142.8 |
21.1 |
5.40 |
3 |
77 |
106.9 |
7.9 |
3.68 |
3 |
170 |
100.3 |
10.4 |
3.96 |
4 |
152 |
-139.4 |
18.3 |
7.79 |
7 |
0.0 |
-116.8 |
12.0 |
5.32 |
LOG |
NP |
1 |
30 |
0.2 |
8.2 |
5.36 |
1 |
86 |
0.1 |
4.1 |
3.08 |
3 |
126 |
-0.5 |
61.5 |
25.18 |
6 |
40 |
0.1 |
5.7 |
4.14 |
1NP = Extra nitrogen + pesticide, P = No extra
nitrogen + pesticide, and O = No extra nitrogen + no pesticide 2Additive
effects = (Prisma - Apex)/2; 3 DTF = days to 50% flower, DMAT
= days to physiological maturity PH = plant height, EARM2 = number of
ears per square metre, KG/HA = grain yield in kg/ha
TKW = weight of a thousand kernels, KE = number of kernels per ear, and
LOG = lodging score
|
In terms of days to 95% physiological maturity (DMAT), five
QTLs were detected in the NP environment on chromosomes 1, 3 (2 QTLs), 4 and
7. Three QTLs were detected in chromosomes 1, 3 and 7, in the P and O environments.
On chromosomes 1 and 3 (position 126 cM) in all the three environments the Prisma
alleles had positive effects on days to 95% physiological maturity and had negative
effects on it on chromosomes 3 (position 145 cM), 4 and 7. In the P and O environments,
the QTL on chromosome 3 (position 126 cM) had large effects on days to 95% physiological
maturity, with phenotypic variation explained by 51.5% and 43.9%, respectively.
The respective additive effects were 1.8 and 1.5 days.
Five QTLs significantly affecting plant height (PH) were detected
on chromosomes 1 (2 QTLs), 3, 4 and 5, in the NP environmemt, seven in the P
environment on chromosomes 1 (2 QTLs), 2 (2 QTLs), 3, 4 and 5. Seven QTLs were
detected in the O environment on chromosomes 1 (3 QTLs) 2, 3, 4 and 5. The Prisma
alleles had positive effects on plant height on chromosomes 1 (position 68 cM),
4 and 5 in all the environments. The Prisma allele had negative effects on plant
height on chromosomes 2 (position 120 cM) and 3. The QTL on chromosome 3 had
a large and dramatic effect on plant height with % variation explained of 73.3%
(NP), 75.1% (P) and 76.3% (O) environments. The respective additive effects
were -9.2, -9.6 and -10.1 cm.
Five QTLs affecting number of ears per square metre were detected
in the NP environment on chromosomes 1, 2, 3, 4 and 6 and five in the P environment
on chromosomes 2, 3, 4 and 7 (2 QTLs) and seven on chromosomes 1, 2, 3, 4, 6
and 7 (2 QTLs) in the O environment. The Prisma alleles had positive effects
on number of ears per square metre on chromosome 3 and 7.
Five QTLs affecting number of kernels per ear were detected
on chromosomes 1, 2, 3, 4 and 7 in the NP environment, six QTLs in the P environment
on chromosomes 2 (2 QTLs), 3 (2 QTLs) 4 and 7 and five in the O environment
on chromosomes 2 (2 QTLs), 3 (2 QTLs) and 7. The Prisma alleles had positive
effects on number of kernels per ear on chromosomes 1, 2 and 4 but had negative
effects on it on chromosomes 3 and 7.
In the case of weight of a thousand kernels (TKW), five QTLs
were mapped on chromosomes 1, 2, 6 and 7 (2 QTLs) in the NP environment, five
in the P environment on chromosomes 1, 2, 3, 4 and seven QTLs on chromosomes
1, 2, 3, 4 (2 QTLs), 6 and 7 in the O environment. On chromosomes 1, 2 and 6
the Prisma alleles had positive effects on weight of a thousand kernels. They
had negative effects on it on chromosomes 3, 4 and 7.
Five significant QTLs located on chromosomes 3 (2 QTLs), 4
and 7 (2 QTLs) affected grain yield in the NP environment. Four QTLs on chromosomes
1, 3, 4 and 6 were detected for grain yield in the P environment and five on
chromosomes 1, 3(2 QTLs), 4 and 7 in the O environment. The Prisma alleles had
positive effects on yield on chromosomes 1 and 3 and had negative effects on
yield on chromosomes 4 and 7 (position 0.0 cM) in all environments.
There was no lodging observed in the P and O environments.
Five QTLs affecting lodging were detected on chromosomes 1 (2 QTLs), 3 (2 QTLs)
and 6. The Prisma allele had negative effects on lodging only on chromosomes
3 (position 126 cM), but positive effects on lodging on all other chromosome
regions. The QTL on chromosome 3 (position 126 cM) had a large effect on lodging.
It explained 61.5% of the phenotypic variation and the additive effect was -0.5.
DISCUSSION
The alpha-lattice design used in this study was efficient,
as there were highly significant blocking effects. As a result, adjusted means,
that is LSMEANS, were used for QTL analyses and estimation of trait correlation
coefficients. The use of means in the estimation of correlation coefficients
is more appropriate than the use of raw data, because this reduces the experimental
error in correlation analyses and hence avoids confounding the residual squares
with the experimental error (Gomez and Gomez, 1984). The significant line effects
shown by the different traits indicated the existence of genetic variation within
the population used in the study.
Significant correlation coefficients among traits indicated
that genes for these traits are either linked, have a pleiotropic effect or
are influenced similarly by the environment (Aasteveit and Aastveit, 1993 quoted
by Veldboom, 1996). The lack of correlation among the traits may suggest that
the loci affecting these traits are independent from each other.
It could be expected that at least some of the genes underlying
QTLs would show QTL x environment interaction, because of the large GxE interaction
observed in all agronomic traits except days to 50% flowering as suggested by
Hayes et al. (1993). QTL x environment interaction would be expressed
as 1) significant effects detected only in a subset of the total number of environment,
in other words, change in the position or disappearing of a QTL due to change
of environment; 2) significant changes in the magnitude of QTL effects across
environments; and, 3) significant opposite favourable alleles of a QTL in distinct
environments otherwise called cross-over interactions (Hayes et al., 1993).
Some of the QTLs, observed in this study, with large effects on traits showed
QTL x environment interaction. This might indicate that genes for the different
traits were being expressed differently in different environments.
Of the eight major yield QTLs identified, only one QTL on chromosome
4 behaved homo-geneously across the three environments. Favourable alleles on
chromosome 4 were contributed by Apex and probably due to its mlo-mutant allele.
This locus was well balanced in terms of many favourable effects. The Apex alleles
positive effects on grain yield and number of ears per square metre is quite
a favourable situation. Both number of ears per square metre and grain yield
were positively correlated. Its positive effect on number of ears per square
metre could have increased grain yield. The locis positive effects on weight
of a thousand kernels in the O environment could have also contributed to high
yields. Weight of a thousand kernels was also earlier on found being positively
correlated with grain yield in this particular environment. Although the QTL
had negative effects on number of kernels per ear in the NP and P environments,
this did not cause significant yield reduction. This is not surprising as the
correlation between number of kernels per ear and yield was not significant.
The QTLs negative effects on number of kernels per ear could have been due
to its large and positive effects on number of ears per square metre, a trait
that was earlier on found to be highly and negatively correlated with it. This
balance of favourable effects showed that this locus was adapted to all the
three environments used and hence showed broad adaptation.
All the other chromosomes had QTLs that were heterogeneously
expressed. Chromosome 1 had a site with large and significant effects on yield
only in the P and O environments. The QTL had positive effects on days to 95%
physiological maturity and days to 50% flower in these environments. The increased
growth duration could have favoured the high yields observed in the two environments.
This QTL had also positive effects on number of ears per square metre in the
O environment, a trait that was found to be highly correlated with yield. However,
the QTL did not have significant effects on yield in the NP environments. For
the NP environment there was a counter balance of favourable and unfavourable
effects. The large plant height could have contributed to lodging score, resulting
in poor light interception and hence low yields. The advantage of having a longer
grain filling duration was negated by lodging score. The positive effects of
the QTL on number of kernels per ear, to have favourable effects on yield could
have been counter balanced by its negative effects on number of ears per square
metre. The locus is not adapted to the NP environment but to the P and O environments.
The QTL on chromosome 3 (the suspected region of the denso
mutant allele) had significant effects on yield in the NP and P environments.
In these environments, this locus showed a balance in some important yield determining
traits. Its positive effects on yield in the NP and P environments could have
been due to its positive effects on days to 50% flower, days to 95% physiological
maturity and number of ears per square metre and its negative effects on plant
height, coupled with low lodging score especially in the NP environment. The
locis negative effects on number of kernels per ear in the NP and number of
kernels per ear and weight of a thousand kernels in the P environments could
have been balanced by its positive effects on number of ears per square metre
which probably compensated for the low number of kernels per ear and weight
of a thousand kernels. This region of the denso mutant allele did not have significant
effects on yield in the O environment. It seemed that, it was no longer an advantage
to have dwarf statured plants in this particular environment. This is expected
as such environments do not favour lodging to have negative effects on yield.
This locus therefore, showed adaptation to the NP and P environments.
The QTLxE interaction was more complex in chromosomes 3 (position
77 cM), 6 and 7 (position 101 cM) as these were characterised by cross-over
of favourable alleles in the different environments in grain yield, although
it was not significant (Lod Score >3). The QTL on chromosome 3 (position
77 cM) had significant positive effects on yield in the O environments but had
negative effects on it, although not significant, in the NP environments. The
QTL on chromosome 6 had significant and positive effects on yield in the P environment
and negative effects on it, even though not significant, in the O environment.
The QTL on chromosome 7 had significant negative effects on yield in the NP
environment and had positive effects on it, although not significant, in the
O environment. The positive effects on yield of the QTL on chromosome 6 in the
P environment could have been due to its positive effects on weight of a thousand
kernels. Its positive effects on lodging score could have caused low yields
in the NP environment. This locus appeared to be more adapted only to the P
environment. The two yield QTLs on chromosomes 7 and 3 (position 170 cM) did
not share their region with other yield related traits. However, because of
their differential expressions across the three environments, the QTLs on chromosome
7 (position 0.0) and 3 (position 170) are more suitable for the NP and O environments.
CONCLUSIONS
The regions of the genome controlling yield and related traits
were differentially expressed across the three environments. This implied that
genes for yield were being expressed differently in different environments.
Breeding programmes designed for marginal areas should therefore be conducted
under those environments.
ACKNOWLEDGEMENTS
This paper is part of an Msc. thesis submitted to the Laboratory
of Plant Breeding, Wageningen Agricultural university. I wish to thank Dr. Johan
Dourleijn and Prof. Piet Stam for supervising this project. Mr Masselink the
farm manager at the university provided valuable technical input in this project.
REFERENCES
- Braun, H.J., Rajaram, S. and van Ginkel, M. 1997. CIMMYTs approach to breeding
for wide adaptation. In: Adaptation in plant breeding. Tigerstedt,
P.M.A.(Eds.), pp.197-205. Kluwer Academic Publishers.
- Ceccarelli, S. 1997. Adaptation to low/high input cultivation. In: Adaptation
in plant breeding. Tigerstedt, P.M.A.(Eds.), pp. 225-236. Kluwer Academic
Publishers.
- Ceccarelli, S. and Hammer, G. L. 1996. Positive interpretation of genotype
by environment interactions in relation to sustainability and biodiversity.
In: Plant adaptation and crop improvement. Cooper, M. and Hammer,
G.L. (Eds.), pp. 467-486.
- Ceccarelli, S., Erskine, W., Hamblin, J. and Grando, S. 1994. Genotype
by environment interaction and international breeding programmes. Experimental
Agriculture 30:177-187.
- Falconer, D.S. 1990. Selection in different environments, effects on environmental
sensitivity (reaction norm) and on mean performance. Genetic Research,
Cambridge, UK. 56-57pp.
- Gomez, K. A. and Gomez, A.A. 1984. Statistical procedures for agricultural
research. Second edition. John Wiley and Sons. New York 680pp.
- Hawtin, G., Iwanaga, M. and Hodgin, T. 1997. Genetic resources in breeding
for adaptation. In: Adaptation in plant breeding. Tigerstedt, P.M.A.
(Ed.), pp. 277-288. Kluwer Academic Publishers.
- Hayes, P.M., Liu, B. H., Knapp, S.J., Chen, F., Jones, B., Blake, T., Franckowiak
J., Rasmusson, D., Sorrells, M., Ullrich, S.E., Wesenberg D. and Kleinhofs,
A. 1993. Quantitative trait locus effects and environmental interaction in
a sample of North American barley germplasm. Theoretical and Applied Genetics
87:392-401.
- Kearsey, M.J. and Pooni, H.S. 1996. The genetical analyses of quantitative
traits. Chapman and Hall, New York, USA.
- Perez-de-la-Vega, M. and Tigerstedt, P.M.A. 1996. Plant genetic adaptedness
to climatic and edaphic environment. Euphytica 92:27-38.
- Snedecor, G.W. and Cochran, W.G. 1989. Statistical methods. Eighth
edition. Iowa State University Press, Ames.
- Veldboom, L.R. and Lee, M. 1996. Genetic mapping of quantitative trait loci
in maize in stress and nonstress environments: II. Plant height and flowering.
Crop Science 36:1320-1327.
|