Foxtail millet (Setaria italica L.) belongs
to Poaceae family (Li 1997) and is one of the most important coarse cereals in
Shanxi Province (Chai and Feng 2003; Liang and Cheng 2005). Foxtail millet has
developed characteristics such as high temperature, drought resistance, barren
tolerance, and strong adaptability (Diao 2011; Jia et al. 2013); thus, it can be planted in various areas of Shanxi
Province. Qin Zhou Huang, Fen Zhou Xiang, and Dong Fang Liang are the brand
name products from foxtail millet which are famous in China. Shanxi is one of
the most important foxtail millet producing provinces in China, where its
cultivation area accounts for about 25% of the total cultivation area of
foxtail millet in China (Gu and Gu 2007; Li et
al. 2015). Therefore, developing new high yielding foxtail millet varieties
suitable to different growing regions along with improving the production level
as a way to promote economic development is of great importance.
Foxtail
millet is the carrier of inheriting Chinese agricultural civilization. A
comprehensive and systematic study on the genetic resources of foxtail millet
in the province - selection of a representative core collection was made from a
large genetic resource collection to carry out a unified and precise phenotypic
identification and a thorough evaluation of the characteristics of local
varieties (Li and Wu 1996) were carried out to cultivate new varieties that
meet the current social needs and improve the use efficiency of local varieties
(Wang et al. 2016).
Germplasm
resources, which are used to develop new varieties provide basic material
for germplasm innovation. The identification of breeding objectives is based on
germplasm
resources; therefore, the collection and preservation of germplasm resources are
important (Zhang 2003). However, the urgent problem is how to study and utilize
large germplasm resources. The concept of core collection provides an efficient
way of making use of large resources (Frankel 1984; Li et al. 1999; Chen et al.
2009). The genetic diversity of foxtail millet germplasm is determined to a
great extent by the minimum size of the foxtail millet germplasm sample, which
provides an effective and practical method for the evaluation of foxtail millet
varieties as well as the promotion of its utilization. Now, more than 27000
foxtail millet germplasm accessions are preserved in the national medium-term
genebanks (Liu et al. 2009), which
account for approximately 80% of the global foxtail millet accessions (Jia and
Diao 2017). Jia et al. (2013)
constructed a core collection from more than 27000 foxtail millet accessions
preserved in the national medium-term genebanks. Wang et al. (2016) identified the comprehensive set of agronomic traits
in core germplasm collections of foxtail millet, which makes it possible to
make efficient use of its gene pool in different ecological areas of China.
At
present, 5627 foxtail millet landraces are
preserved in the seed bank of the Shanxi Province (Wen et al. 2002), among which a large number have well-developed traits
such as drought resistance, increasing yield potential, resistance to smut,
excellent aroma and the good eating quality. Wang et al. (2019) made use of 5627 landraces from Shanxi Province as
source material
and constructed a primary core collection of foxtail millet landraces. In
the present study, the material included a foxtail millet core collection
comprising of 595 accessions. Comprehensive phenotypic identification of the
accessions was carried out, which laid the foundation for further understanding
of the phenotypic diversity of a core collection of foxtail millet landraces.
The objective of this study is to characterize core collection of foxtail
millet using morphological marker and use comprehensive information to evaluate
each germplasm, and also for the efficient use of these accessions in future foxtail millet improvement program.
Materials and Methods
Plant materials
For two consecutive years (2017–2018), 18 traits in 595 foxtail millet
accessions were investigated at the Dongyang experimental base in Shanxi
Academy of Agricultural Sciences (37.6 °N, 112.7 °E). The experimental site has
a temperate continental climate, with the annual average temperature of 9.8°C
and annual average of 450.5 mm. All accessions were planted on 18th
May 2017 and 27th May 2018. Two rows of 3 m long of each genotype
were planted with the row- to- row spacing of 35 cm and the plant- to- plant
spacing of 10 cm. The experimental soil was red clay loam having pH of 8.1 and medium in organic material
(18.2 g kg-1). Fertilizers were applied at the rate of 66:36:54kg/ha
N:P2O5:K2O as basal dose during land
preparation and 30 kg/ha N was top-dressed 50 days after seeding. The test was
irrigated once before heading date. All germplasms were provided by the
Institute of Crop Germplasm Resources, Shanxi Academy of Agricultural Sciences.
Investigation of traits
The data were recorded as per the
descriptors of foxtail millet [Setaria
italica (L.) beauv.] (Lu 2006). Phenotypic traits including sheath
color, tillering, panicle type, heading date, stem node number,
plant height, peduncle length, leaf length, leaf width,
panicle length, panicle diameter, diameter of main stem, panicle weight per
main stem, grain weight per main stem, primary branch number
per panicle, spikelet number per primary branch, grain weight
(1000-seed) and protein content were investigated.
Data analysis
The phenotypic data were entered
and analyzed using Microsoft Excel 2007. The
fuzzy membership function values for each trait were computed using a
membership function whose range is between 0 and
1. The formula is as follows:
µ (xi) = (xi-ximin)/(ximax-ximin)
(i=1, 2, 3,……, 595)
µ (xi) is the membership
function value of the personality trait of one material, xi is the
personality value, and ximax and ximin
are the maximum and minimum values of xi, respectively.
Shannon-Wiener
diversity index (H') was used
to evaluate genetic diversity. The formula is as follows:
H'= –∑ (pi Ln pi) (i=1, 2, 3,……,
595)
Data analysis was carried out using
correlation analysis, principal component analysis (PCA) and cluster analysis
on SPSS 19.5. The scores of each principal component were calculated using
Microsoft Excel 2007. The evaluation indices of all traits in foxtail millet
germplasm accessions were screened using the stepwise regression
analysis.
Results
Genetic diversity of qualitative traits
There were significant differences in the qualitative traits of the core
collection of the foxtail millet germplasm in
Shanxi Province (Fig. 1). Out of 595 genotypes, the majority were found to have
green leaf sheath color (56.66%), while 26.64% of genotypes had a purple color.
With regard to tillering capacity, 53.38% of foxtail millet genotypes had
tillers and others had no-tillers. In terms of panicle type (Fig. 2), most of
the genotypes belonged to type 2, accounting for 72.25%, followed by type 3
(13.87%), type 6 (6.6%) and type 4 (6.4%).
Genetic diversity of quantitative traits
According to the classification of
genetic diversity, the quantitative traits were distributed into 10 groups
(Fig. 3). The panicle diameter and grain weight (1000-seed) had values
between 0.4 and 0.8 on a 10 point (0.1–1) scale, which represented
91.4% and 91.9% of the total genetic diversity, respectively. Furthermore, the
diameter of the main stem, primary branch number per panicle and spikelet
number per primary branch were in the range of 0.3–0.6, accounting for 96.3%,
86.6%, and 94.3%, respectively. Other quantitative traits showed the values
between 0.3–0.7, representing 87.4% to 95.6% of the diversity.
%20405-412_files/image002.png)
Fig. 1: The number
of different types of germplasm in leaf sheath color, tillering capacity and
panicle type in 595 core collection
%20405-412_files/image004.jpg)
Fig. 2: Diversity
panicle types of Shanxi foxtail millet core collection
%20405-412_files/image006.png)
Fig. 3: Distribution diagram of subordinate function value of fifteen phenotypic traits
The
variation and distribution of phenotypic traits in the core collection of
foxtail millet are presented in Table 1. The minimum coefficient of variation
among 15 quantitative traits was for protein content (6.71), while the maximum
was for the grain weight per main stem (26.28). The minimum (1.29) and maximum
(1.95) genetic diversity index values were observed for the diameter of the
main stem and the protein content, respectively.
The
average heading date (HD) was 62.48 ± 5.88 and ranged from 46.0 to 86.5 d.
Among them, H526 (Majiangsheng 00018180) had the longest heading date (86.5),
while the shortest heading date (46.0) was observed in H087 (Sangenqimaogu
00005564).
Table 1: Phenotypic traits of 595 foxtail millet
accessions
|
Agronomic
trait |
Mean ±SD |
Range |
CV (%) |
H' |
|
HD (d) |
62.48 ±5.88 |
46.00–86.5 |
9.42 |
1.81 |
|
SN |
12.02 ±1.28 |
5.10–16.1 |
10.68 |
1.58 |
|
PH (cm) |
114.55 ±17.04 |
38.90–160 |
14.88 |
1.74 |
|
PeL (cm) |
38.42 ±6.61 |
20.20–59.7 |
17.20 |
1.93 |
|
LL (cm) |
44.46 ±4.94 |
29.00–62.8 |
11.10 |
1.81 |
|
LW (cm) |
2.81 ±0.25 |
1.88–4.47 |
8.84 |
1.35 |
|
PL (cm) |
28.04 ±4.80 |
14.20–44.5 |
17.11 |
1.88 |
|
PD (cm) |
2.51 ±0.35 |
1.22–3.68 |
14.11 |
1.79 |
|
DMS (cm) |
0.80 ±0.08 |
0.46–1.36 |
10.39 |
1.29 |
|
PW (g) |
24.33 ±6.08 |
5.24–46.84 |
25.00 |
1.81 |
|
GW (g) |
19.48 ±5.12 |
3.67–37.11 |
26.28 |
1.84 |
|
PBNP |
112.88 ±21.06 |
54.10–194.6 |
18.66 |
1.78 |
|
SNPB |
90.52 ±19.30 |
19.50–168.1 |
21.33 |
1.69 |
|
SW (g) |
3.15 ±0.36 |
1.63–4.19 |
11.28 |
1.75 |
|
PC (%) |
12.37 ±0.83 |
10.00–14.8 |
6.71 |
1.95 |
HD, heading
date; SN, stem node number; PH, plant height; PeL, peduncle length; LL, leaf
length; LW, leaf width; PL, panicle length; PD, panicle diameter; DMS, diameter
of main stem; PW, panicle weight per main stem;
GW, grain weight per main stem; PBNP, primary branch number per panicle; SNPB,
spikelet number per primary branch; SW, 1000-seed weight; PC, protein content
Panicle and grain weight per main stem are the
two important yield component traits of foxtail millet. The average panicle
weight was 24.33 ±6.08 g, ranging from 5.24 to
46.84 g. The average grain weight obtained was 19.48 ±5.12 g, with range values
of 3.67 –37.11 g. Among them, H585 (Dagu 00016114) and H252 (Baimaolianggu
00015626) had the lowest panicle and grain weight, while the highest panicle
and grain weights were recorded in H180 (Zhuhanchangruangu 00006232) and H266
(Jinjunqu 00015736).
The grain
weight (1000-seed) ranged from 1.63–4.19 g, with a mean value of 3.15 ±0.36 g.
Of the genotypes, H252 (Baigangu 00015626) had the lowest (1.63 g) and H268
(Huangruangu 00015749) had the highest (4.19 g) 1000-grain weight. The protein
content had an average value of 12.37 ±0.83%, with a range of 10.01– 14.81%.
The lowest (10.01%) and highest (14.81%) protein contents were found in H031
(Dahonghuanhgu 0004898) and H291 (Huangruangu 00015945), respectively.
Correlation analysis of
phenotypic traits
Correlation analysis of the studied
traits showed that there were different degrees of correlations among the 18
phenotypic traits (Fig. 4). The correlation between qualitative and quantitative
traits was weak. Heading date was positively correlated with SN, PH, DMS, and
PBNP, indicating the positive effect of HD on the growth of foxtail Table 2: The variation distribution of quantitative traits in different years
|
Trait |
2017 |
2018 |
||
|
Mean |
CV |
Mean |
CV |
|
|
62.42 ±8.72 |
13.97 |
62.54 ±4.97 |
7.95 |
|
|
SN |
12.03 ±1.46 |
12.14 |
12.02 ±1.33 |
11.06 |
|
PH |
109.82 ±17.33 |
15.78 |
119.38 ±18.69 |
15.66 |
|
PeL |
40.48 ±7.88 |
19.46 |
36.31 ±6.54 |
18.02 |
|
LL |
44.75 ±5.83 |
13.04 |
44.2 ±5.41 |
12.24 |
|
LW |
2.97 ±0.32 |
10.85 |
2.66 ±0.28 |
10.48 |
|
PL |
27.00 ±4.94 |
18.31 |
29.10 ±5.67 |
19.49 |
|
PD |
2.54 ±0.4 |
15.74 |
2.48 ±0.44 |
17.59 |
|
DMS |
0.77 ±0.1 |
13.69 |
0.83 ±0.1 |
12.47 |
|
PW |
25.78 ±8.04 |
31.18 |
22.86 ±6.69 |
29.25 |
|
GW |
20.04 ±6.62 |
33.06 |
18.92 ±5.78 |
30.56 |
|
PBNP |
108.95 ±24.67 |
22.65 |
117.10 ±21.76 |
18.59 |
|
SNPB |
97.10 ±25.52 |
26.29 |
83.76 ±21.18 |
25.29 |
|
SW |
3.08 ±0.45 |
14.67 |
3.24 ±0.4 |
12.26 |
|
PC |
12.64 ±1.19 |
9.41 |
12.10 ±0.79 |
6.57 |
%20405-412_files/image008.jpg)
Fig. 4: Correlation matrix between the agronomic traits. Trait codes are given in Supplementary Table 1
millet. The yield component traits
PW and GW had highly significant positive correlations with the growth indices
such as SN, PH, LL, LW, PL, PD, and DMS. It was found that SNPB and SW were
also positively correlated with these traits, indicating the close interaction
between vegetative and reproductive growth during the foxtail millet
development. The yield of foxtail millet could only be improved by coordinating
the vegetative and reproductive growth. The PC was significantly positively
correlated with PeL and had a negative correlation with other traits except for
SW and SNPB. Therefore, the use of breeding processes to improve foxtail millet
protein quality is of significant importance.
Annual variation in
phenotypic traits
Stability is an important component
of germplasm resources. Due to the differences in ecological environments,
there might be differences in phenotypic traits in foxtail millet during two
consecutive years. Table 2 lists the mean (average) and coefficient of
variation (CV) for phenotypic traits in
foxtail millet for two years. The mean values of HD, SN, LL, LW, and PD showed
small variations between the two years. The highest mean values of PH, PL, DMS,
PBNP, and SW were obtained in 2018, while other quantitative traits had higher
values in 2017. There were obvious variations in PL, PW, GW, and SNPB, which
were significantly related to yield composition. Except for PL and PD, the
coefficients of variation for other quantitative traits were higher in 2017
than those in 2018 and obvious variations were observed for HD, PBNP, SW, and
PC.
Comparisons of shannon-wiener diversity indices of 18
phenotypic traits in core collection
According to administrative
divisions in Shanxi Province, core collections of foxtail millet genetic
resources, representing 11 cities, could be divided into 11 categories, namely,
Datong (DT), Shuozhou (SZ), Xinzhou (XZ), Taiyuan (TY), Jinzhong (JZ), Lvliang
(LL), Yangquan (YQ), Changzhi (CZ), Linfen (LF), Jincheng (JC) and Yuncheng
(YC). Statistical analysis of genetic diversity indices of 18 phenotypic traits
in a core collection consisting of 595 accessions was done based on 11 cities.
There were small differences among cities for the mean
values of diversity indices (Table 3), indicating
that phenotypic diversity in each city was equally abundant. However, the genetic
diversity index (H') had the lowest mean value (1.6528) for Jincheng city and
the highest value (1.8725) for Taiyuan city. Genetic diversity indices of the
qualitative traits were significantly lower than that of the quantitative
traits. The H' values of three quality traits (SC, TC and PT) were lower.
Cluster analysis on samplings of different regions in shanxi province
The core collection was divided into three groups
with the Euclidean distance of 5.39 (Fig. 5) and the results were analyzed by
DPS software. Datong and Shuozhou could be classified under Cluster I, while Cluster
II included Jincheng, Jinzhong, Taiyuan, Luliang, Xinzhou, Yangquan,
Changzhi and Linfen, and Yuncheng belonged
to cluster III.
Principal component analysis of phenotypic traits
Correlation analysis showed
correlations among phenotypic traits, which had a certain effect on the
evaluation of the core collection. In order to mitigate the adverse effect of
the related factors, the core collection was comprehensively evaluated by
principal component analysis using DPS software. The cumulative contribution
rate of the first nine principal components is over 80%, indicating that these
components represented more than 80% of the genetic information for the
evaluation of core collections.
The contribution rate of the first principal component
was 25.71% and the power vectors of PW and GW were larger than for other
traits, thus the first principal component was composed of these two traits
(Table 4). The contribution rate of the second principal component was 14.26%,
which was dominated by HD, SN, and PH, indicating that the second principal
component was a comprehensive reflection of these traits. The contribution rate
of the third principal component was 11.18%, and the power vectors of LL, PL,
DP, PBNP, and SNPB were larger than for other characters, which showed that the
third principal component consisted of these characters. The contribution rate
of the fourth component was 6.42%, and LW and DMS had larger power vectors than
that of other characters. The fifth principal component had a contribution rate
of 6.13% and it was composed of PT. The
%20405-412_files/image010.jpg)
Fig. 5: Clustering dendrogram of foxtail millet germplasm resources from different regions in Shanxi Province
contribution rate of the sixth principal component
was 5.44%, and the power vector of SC was the highest. The contribution rate of
the seventh principal component was 4.91%, and the PC value was the largest.
Moreover, the contribution rate of the eighth principal component was 4.58%,
and the power vector of TC was the largest. The contribution rate of the ninth
principal component was 4.40%, and PeL and SW had the highest values,
indicating that these two traits had the largest contribution to this
component.
Table 3: Comparison
of the Shannon-Wiener diversity indexes of the 18 phenotypic traits of foxtail
millet core collection
|
|
SC |
TC |
PT |
HD |
SN |
PH |
PeL |
LL |
LW |
PL |
PD |
DMS |
PW |
GW |
PBNP |
SNPB |
SW |
PC |
Mean |
|
DT |
1.034 |
0.621 |
1.143 |
1.971 |
1.881 |
1.797 |
1.981 |
1.907 |
2.067 |
1.813 |
1.728 |
1.912 |
1.955 |
1.918 |
1.812 |
1.879 |
1.750 |
1.927 |
1.727 |
|
SZ |
0.974 |
0.562 |
1.180 |
1.927 |
1.927 |
1.808 |
1.652 |
1.771 |
1.841 |
2.047 |
2.068 |
1.602 |
2.220 |
2.047 |
1.841 |
1.836 |
1.820 |
1.927 |
1.725 |
|
XZ |
0.954 |
0.609 |
0.815 |
2.076 |
1.972 |
2.148 |
2.107 |
2.030 |
2.159 |
1.784 |
2.098 |
1.708 |
2.019 |
2.092 |
2.048 |
1.904 |
1.866 |
1.961 |
1.797 |
|
TY |
0.957 |
0.666 |
0.536 |
2.109 |
2.021 |
2.172 |
2.122 |
2.025 |
2.139 |
2.162 |
2.139 |
1.998 |
2.144 |
2.264 |
2.055 |
1.945 |
2.142 |
2.107 |
1.873 |
|
JZ |
0.959 |
0.693 |
0.713 |
1.908 |
1.984 |
1.872 |
2.069 |
1.801 |
1.752 |
2.012 |
1.971 |
1.769 |
2.094 |
2.061 |
1.959 |
1.931 |
1.901 |
1.942 |
1.744 |
|
LL |
0.869 |
0.607 |
0.842 |
2.198 |
1.587 |
1.747 |
2.027 |
2.082 |
1.923 |
2.000 |
2.150 |
1.333 |
1.962 |
2.006 |
2.037 |
1.869 |
1.820 |
2.095 |
1.731 |
|
YQ |
1.028 |
0.649 |
0.703 |
1.956 |
1.844 |
1.921 |
1.905 |
1.891 |
2.201 |
1.787 |
1.624 |
1.844 |
2.038 |
1.956 |
1.824 |
2.069 |
2.089 |
2.150 |
1.749 |
|
CZ |
0.906 |
0.640 |
0.864 |
2.069 |
2.051 |
2.034 |
1.934 |
2.044 |
1.877 |
2.049 |
2.123 |
1.667 |
2.060 |
2.082 |
1.662 |
1.758 |
1.991 |
2.154 |
1.776 |
|
LF |
1.020 |
0.693 |
0.933 |
1.909 |
2.143 |
1.927 |
2.079 |
2.075 |
1.498 |
1.724 |
2.078 |
1.730 |
2.227 |
2.192 |
2.083 |
2.091 |
2.105 |
2.049 |
1.808 |
|
JC |
0.451 |
0.637 |
0.824 |
1.676 |
1.979 |
1.864 |
1.705 |
1.792 |
1.864 |
1.907 |
1.979 |
1.864 |
1.864 |
1.820 |
1.979 |
1.864 |
1.864 |
1.820 |
1.653 |
|
YC |
0.831 |
0.257 |
0.628 |
1.710 |
1.767 |
1.615 |
2.061 |
1.864 |
1.841 |
1.991 |
1.859 |
1.807 |
1.991 |
2.119 |
1.841 |
2.084 |
2.084 |
1.859 |
1.678 |
|
Mean |
0.973 |
0.691 |
0.914 |
1.806 |
1.579 |
1.738 |
1.926 |
1.811 |
1.348 |
1.877 |
1.791 |
1.289 |
1.808 |
1.841 |
1.777 |
1.670 |
1.753 |
1.948 |
|
Trait codes are given in Table 1
Table 4: Power vector (PV), eigenvalues
(E), contribution rate (CR), and cumulative contribution rate (CCR) of the
first nine principal components based on 18 phenotypic traits
|
Trait |
PV1 |
PV2 |
PV3 |
PV4 |
PV5 |
PV6 |
PV7 |
PV8 |
PV9 |
|
SC |
0.0034 |
-0.1620 |
-0.1186 |
0.4297 |
0.2511 |
-0.6200 |
0.0422 |
-0.1491 |
0.4979 |
|
TC |
0.0803 |
0.2654 |
0.0616 |
0.2047 |
-0.0235 |
-0.2523 |
-0.4248 |
0.7754 |
-0.0345 |
|
PT |
-0.0418 |
-0.0753 |
0.0495 |
0.3612 |
0.7277 |
0.1974 |
-0.0378 |
-0.0547 |
-0.4042 |
|
HD |
0.1406 |
-0.4161 |
0.1691 |
0.0802 |
-0.1019 |
0.1378 |
-0.1108 |
0.1861 |
-0.1736 |
|
SN |
0.2802 |
-0.4046 |
-0.0160 |
-0.1129 |
0.0704 |
0.0866 |
-0.1436 |
0.0793 |
0.1624 |
|
PH |
0.2659 |
-0.3902 |
0.0577 |
-0.1372 |
0.1564 |
0.1006 |
-0.1853 |
-0.0292 |
0.2727 |
|
PeL |
0.0731 |
0.3340 |
0.2763 |
-0.0315 |
0.1557 |
0.3196 |
-0.3450 |
-0.1533 |
0.3670 |
|
LL |
0.2921 |
0.1470 |
0.3749 |
0.0640 |
0.0392 |
0.1086 |
-0.0183 |
-0.0974 |
0.1611 |
|
LW |
0.2471 |
0.2181 |
-0.1317 |
0.3854 |
-0.2431 |
0.1158 |
0.2518 |
-0.0944 |
-0.0280 |
|
PL |
0.2355 |
0.2317 |
0.4194 |
0.0185 |
-0.0288 |
-0.0584 |
0.1153 |
-0.0638 |
0.0896 |
|
DP |
0.2249 |
0.1090 |
-0.3332 |
0.2833 |
-0.2139 |
0.1815 |
-0.0914 |
-0.1321 |
0.1180 |
|
DMS |
0.2728 |
-0.0861 |
-0.0296 |
0.4180 |
-0.0472 |
0.2606 |
0.1578 |
0.1157 |
-0.0835 |
|
PW |
0.4086 |
0.0904 |
-0.1237 |
-0.1590 |
0.1137 |
-0.1444 |
0.1841 |
0.0554 |
-0.1419 |
|
GW |
0.4012 |
0.0837 |
-0.1584 |
-0.1930 |
0.1505 |
-0.1518 |
0.1612 |
0.0508 |
-0.1452 |
|
PBNP |
0.1138 |
-0.1757 |
0.4861 |
-0.0344 |
-0.1139 |
-0.2137 |
0.4167 |
0.1164 |
-0.0562 |
|
SNPB |
0.2590 |
0.0605 |
-0.3844 |
-0.3072 |
0.1401 |
0.1110 |
0.0535 |
0.1751 |
0.1798 |
|
SW |
0.2255 |
0.2240 |
0.0318 |
-0.1547 |
0.1255 |
-0.3397 |
-0.2893 |
-0.3479 |
-0.3495 |
|
PC |
-0.1786 |
0.2342 |
-0.0069 |
-0.1183 |
0.3865 |
0.1656 |
0.4469 |
0.2897 |
0.2597 |
|
E |
4.6285 |
2.5671 |
2.0125 |
1.1554 |
1.1041 |
0.9794 |
0.8834 |
0.8254 |
0.7912 |
|
CR% |
25.7140 |
14.2619 |
11.1806 |
6.4189 |
6.1338 |
5.4410 |
4.9079 |
4.5856 |
4.3956 |
|
CCR% |
25.7140 |
39.9759 |
51.1565 |
57.5754 |
63.7092 |
69.1502 |
74.0581 |
78.6437 |
83.0393 |
The
trait codes are listed in Table 1
Comprehensive evaluation of
phenotypic traits
The values of 18 phenotypic traits
of core collection in foxtail millet were standardized and then replaced with
the 9 principal components. The scores of 9 principal components were obtained
from accessions and standardized using the fuzzy membership function. The
weight coefficients of each principal component were calculated (0.310, 0.172,
0.135, 0.077, 0.074, 0.066, 0.059, 0.055, 0.053). Furthermore, the composite
score (F) for each germplasm was calculated. The higher the F value for each
millet germplasm, the better the phenotypic character is. The average F value
for the millet core collection was
0.5227 and the F value of H180 (Zhuhanchangruangu 00006232) from Lishi, Lvliang
City was the highest (0.7235), while H256 (Dahegu 00015665) from Tunliu,
Changzhi City had the lowest F value (0.3127), indicating more comprehensive
phenotypes of H180 and less of H256 (Table 5). Correlation analysis showed that
F value was positively correlated with all phenotypic traits except PT and HD
(Table 6).
Table 5: The higher and lower F value of
accession
|
location |
Lower
F |
Name |
Location |
Higher
F |
|
|
Daheigu |
Yanggao |
0.3127 |
Langweiba |
Zuoquan |
0.6468 |
|
Gegu |
Pingyao |
0.3225 |
Baidihuang |
Xingxian |
0.6487 |
|
Bayihuang |
Tunliu |
0.3319 |
Machangjiang |
Qinyuan |
0.6542 |
|
Qinggu |
Tunliu |
0.3331 |
Dabaigu |
Puxian |
0.6564 |
|
Zhuazhuagu |
Xiaxian |
0.3343 |
Zhushugu |
Yushe |
0.6573 |
|
Huangruangu |
Yingxian |
0.3350 |
Hongxiaoweigu |
Shouyang |
0.6636 |
|
Yanggu |
Pinglu |
0.3413 |
Zigangu |
Pingyao |
0.6640 |
|
Honggangu |
Yanhu |
0.3440 |
Sanbianlian |
Zuoquan |
0.6663 |
|
Xiaobaigu |
Zhangzi |
0.3449 |
Gouchangruangu |
Pingyao |
0.6881 |
|
Qishiliugu |
Pingding |
0.3456 |
Zhuhanchangruangu |
Lishan |
0.7235 |
Table 6: Correlation coefficients between
the 18 phenotypic traits and comprehensive value (F-value)
|
Trait |
Correlation
coefficient |
Trait |
Correlation
coefficient |
|
SC |
-0.131** |
PL |
0.752** |
|
TC |
0.407** |
DP |
0.312** |
|
PT |
0.016 |
DMS |
0.470** |
|
HD |
-0.013 |
PW |
0.637** |
|
SN |
0.138** |
GW |
0.599** |
|
PH |
0.165** |
PBNP |
0.272** |
|
PeL |
0.554** |
SNPB |
0.320** |
|
LL |
0.784** |
SW |
0.358** |
|
LW |
0.555** |
PC |
0.084* |
The
optimal regression equation was constructed based on the composite score values
(F) and phenotypic traits and the evaluation index values for traits of millet
resources were screened. The equation constructed using the stepwise regression
is as follows: y = (–368.446 + 20.610x2 + 2.191x7 +
3.694x8 + 42.224x9 + 3.340x10 + 147.468x12
+ 2.796x13 + 17.301x18) × 10 –3, where x2,
x7, x8, x9, x10, x12, x13
and x18 are representative of TC, PeL, LL, LW, PL, DMS, PW and PC,
respectively and their direct path coefficient were 0.163, 0.229, 0.288, 0.166,
0.254, 0.194, 0.269 and 0.227, respectively. The determination coefficient (R2)
for the above equation was 0.984, indicating 98.4%
of the variation in these 8 independent variables. The total
variation of F was 4396.07 and various coefficients in the equation were highly
significant. The F values for each of the variables in the regression equation
showed high efficiency of 8 variables and the value of the phenotypic variation
in foxtail millet landraces from Shanxi Province was significant. It was
indicated that 8 variables could be used as the core collection comprehensive
evaluation indices.
Discussion
The genetic diversity for landraces
of foxtail millet is required to evolve in response to complex geographical and
ecological variations, climate change and long-term artificial domestication.
These varieties are mainly planted in small regions and may be outdated or
sporadic, but they have different origins and characteristics, possessed high
genetic diversity and are suited to traditional agricultural practices (Li and
Wu 1996; Nakayama 2011). A majority of these resources has not been genetically
improved and may have obvious shortcomings, but some unusual characteristics
are of great importance in breeding (Wang et
al. 2018). The phenotypic traits of 595 core collection were evaluated for
two consecutive years in order to better understand morphological
characteristics of germplasm resources in crop breeding.
Genetic
variation in the core collection is abundant, but it still is lower than the
diversity of foxtail millet core collection at the national level (Wang et al. 2016). The range and coefficients
of variation for all traits were high. The traits such as plant height, panicle
length, panicle weight and grain weight in foxtail millet core collections
showed generally higher values compared to those obtained in the studies by
Tian (2010) and Qu et al. (2018).
However, it was lower than that reported by Wang et al. (2009) on the germplasm identification of foxtail millet in
Gansu Province. In this study, the average coefficient of variation for 15
quantitative traits was 14.87% and the average genetic diversity index was
1.73. Coefficients of variation for panicle weight and grain weight were 25.00
and 26.28, respectively, and their genetic diversity indices were 1.81 and
1.84, respectively, indicating that core collections had abundant phenotypic
diversity and could well represent the diversity in the entire collection of
foxtail millet in Shanxi Province. Genetic diversity of different types of
germplasm provides opportunities in plant breeding. In the present study, the
range of variation for panicle weight, grain weight, 1000-seed weight and
protein content could contain analysis results of
the study by Zhang et al. (2017);
Moreover, the range of variation for panicle length, panicle weight, grain
weight and 1000-seed weight could contain the analysis results of the study by
Xiang et al. (2018), indicating that
gaining the advantages of landraces and improving the existing main varieties
are positive approaches in foxtail millet breeding in Shanxi Province.
The
resources of 11 cities from the same geographical origin in Shanxi Province
were divided into three categories based on the systematic and cluster
sampling. The first category included Datong and Shuozhou, in the north of
Shanxi, which was mainly the early maturing spring- sowing areas; the second
category included the majority of the cities in the south-central part of
Shanxi Province, which were mainly the middle and late-maturing spring-sowing
regions. Moreover, the third category included Yuncheng City, in the south of
Shanxi, which was the summer-sowing area. This classification basically
reflects the geographical and ecological areas of Shanxi for foxtail millet
production.
In this study, the principal component analysis (PCA) for
18 traits of Shanxi core collection in foxtail millet was carried out. After
the normalization of principal component scores, the composite scores (F) were
calculated and the comprehensive performance of each landrace was evaluated and
compared based on a quantitative value. Among the landraces, Langweiba,
Baidihuang, Machangjiang, Dabaigu, Zhushugu, Hongxiaoweigu, Zigangu,
Sanbianlian, Gouchangruangu, and Zhuhanchangruangu showed better comprehensive
performance. The comprehensive phenotypic score of Jingu 21 (00024668) (Chen et al. 1992) was only 0.4697 and it
ranked 494th. The annual cultivated area in Shanxi was about 1.33×105
hm2, which was accepted by farmers and the markets, indicating that
the local varieties still maintain allelic variation that needs to be further
excavated and utilized.
All 16
traits were significantly correlated with F value. The correlation coefficient
between tillering capacity and F value was 0.407 and that of peduncle length,
leaf length, leaf width, panicle length, diameter of main stem, panicle weight,
and grain weight was relatively high. The higher the comprehensive value, the
higher the trait value was. For identification purposes and breeding practices,
close attention should be paid to the selection of these traits. The research
study of genetic variation in a population, combined with F value, can be
conducted.
The correlation analysis showed that there was a
correlation between different traits; although some were low, they were still
significant. Meanwhile, the correlations between qualitative and quantitative
traits were weak or did not exist, which were similar to those found between
phenotypic traits and comprehensive score values. With regard to yield
components panicle weight, grain weight, and 1000-seed weight, the correlation
coefficient was highly significant. In order to improve the yield, the foxtail
millet breeding programs and the selection of main traits should be
strengthened; however, other secondary factors should also be considered.
Varieties whose traits are coordinated and have excellent comprehensive scores
could be extensively cultivated (Lu et
al. 2015; Zhang et al. 2017).
Conclusion
Among the 595 accessions of foxtail millet core
collection in Shanxi province, high level of diversity was observed for 18
agronomic traits like sheath color, tillering capacity, panicle type, heading
date, stem node number, plant height, peduncle length, leaf length and width,
panicle length and diameter, panicle and grain weight per main stem, protein
content, etc. The protein content was significantly positive correlated with
peduncle length. Landraces such as Zhuhanchangruangu, Sanbianlian, Zigangu,
Dabaigu, Machangjiang, Langweiba, etc. had the much better comprehensive traits
important and can usesd for crop improvement programmes. According to the
established regression equation, eight traits were tested as the key phenotypic
indicators. The abundant genetic variation of core collection provides utility
resources for genetic improvement of foxtail millet.
Acknowledgments
The Specialty Industry for Key Research and Development
Program in Shanxi Academy of Agricultural Sciences (YCX2019T01); Key Research
and Development Program of Shanxi Province (201803D221012–7); The Earmarked
Fund for Scientific and Technological Innovation Project for Excellent Talent
of Platform Base and Talents(201705D211026); The Earmarked Fund for China
Agriculture Research System (CARS–06–13.5-A16)
References
Chai Y, BL Feng (2003). Present situation
and developing strategies of minor grain crops in China. Agric Res Arid
Areas 21:145‒151
Chen Y, TY Wei, XL Jing, XX Zhang (1992). Breeding
and popularization prospect of Jingu 21. Shanxi Agric Sci 12:7‒8
Chen Y, DJ Pan, QW Yang, B Liu, ZL Fan, JY Chen, C
Li (2009). Sampling strategy for an applied core collection of gaozhou wildrice
(Oryza rufipogon Griff.) in
Guangdong, China. Acta Agron Sin 35:459‒466
Diao XM (2011). Industry and developmental perspectives
of Chinese foxtail millet//Chinese Industry and Technical System of Foxtail
Millet, pp: 20‒30. Diao XM (ed.). China Agricultural Sciences and Technology Press,
Beijing, China
Frankel OH (1984). Genetic perspectives of
germplasm conservation. In: Genetic Manipulation: Impact on Man and
Society, pp:161‒170. Arber W, K Llimensee, WJ Peacock, P Starlinger (Eds.). Cambridge
University Press, UK
Gu GM, SL Gu (2007). The Origin and Development of Foxtail Millet in Shanxi Province. Chinese Agricultural Science Press, Beijing,
China
Jia GQ, XH Huang, H Zhi, Y Zhao, Q Zhao, WJ Li, Y
Chai, LF Yang, KY Liu, HY Lu, CR Zhu, YQ Lu, CC Zhou, DL Fan, QJ Weng, YL Guo,
T Huang, L Zhang, TT Lu, Q Feng, HF Hao, HK Liu, P Lu, N Zhang, YH Li, EH Guo,
SJ Wang, SY Wang, JR Liu, WF Zhang, GQ Chen, BJ Zhang, W Li, YF Wang, HQ Li, BH
Zhao, JY Li, XM Diao, B Han (2013). A haplotype map of genomic variations and
genome-wide association studies of agronomic traits in foxtail millet (Setaria italica). Nat Genet 45:957‒961
Jia GQ, XM Diao (2017). Current status and
perspectives of researches on foxtail millet (Setaria italica (L.) P. Beauv.): A potential model of plant
functional genomics studies. Chin Bull Life Sci 29:292‒301
Li JB, SG Hou, CP Wang, XF Hou (2015). Thinking on
the production and industry development of millet in Shanxi Province. J
Shanxi Agric Sci 43:1508‒1510
Li Y, SZ Wu (1996). Traditional maintenance and
multiplication of foxtail millet (Setaria
italic (L.) P. Beauv.) landraces in China. Euphytica 87:33‒38
Li YM (1997). Foxtail
Millet Thremmatology. Chinese
Agriculture Press, Beijing, China
Li ZC, HL Zhang, CQ Sun, XS Wang (1999). Status
and prospects of core collection in plant germplasm resource. J Chin Agric
Univ 4:51‒62
Liang SB, RH Cheng (2005). Position of foxtail
millet and small grain the cropping pattern adjustment and its development
strategy in China. J Hebei Agric Sci 9:93‒95
Liu X, Y Li, YS Cao, YS Dong, W Fang, P Lu (2009).
Geographic distribution and germplasm-rich region of cereals in China. J
Plant Genet Resour 10:1‒8
Lu P 2006. Specifications
and Data Standards for Millet Germplasm Resources. Chinese
Agriculture Press, Beijing, China
Lu WT, H Yu, SN Cao, CQ Chen (2015). Effects of
climate warming on growth process and yield of summer maize in Huang-Huai-Hai
plain in last 20 years. Sci Agric Sin 48:3132‒3145
Nakayama H (2011). Genetic variations of foxtail
millet, Setaria italica (L.) P.
Beauv. and their geographical distribution. Bull Nat Insti Agrobiol Resour
73‒144
Qu Y, H Song, Y Liu, KZ Wang, XL Gao, JF Gao, PK
Wang, P Yang, BL Fen (2018). Genetic diversity of agronomic traits and node
characters in foxtail millet (Setaria
italica Beauv.). Agric Res Arid
Areas 36:51‒58
Tian BH (2010). Genetic diversity of landrace and
improved cultivars in foxtail millet. J Plant Genet Resour 11:224‒228
Wang HG, GQ Jia, H Zhi, QF Wen, JL Dong, L Chen,
JJ Wang, XN Cao, SC Liu, L Wang, ZJ Qiao, XM Diao (2016). Phenotypic diversity
evaluations of foxtail millet core collections. Acta Agron Sin 42:19‒30
Wang HG, HB Qin, X Tian, JZ Lv, L Chen, JJ Wang,
XN Cao, SC Liu, L Wang, QF Wen, ZX Mu, ZJ Qiao (2018). Phenotypic diversity of
local foxtail millet varieties in Shanxi. Chin Agric Sci Bull 34:19‒24
Wang HG, QF Wen, ZJ Qiao, ZX Mu (2019). Core
germplasm construction of foxtail millet landrace in Shanxi. J Agron
9:26‒31
Wang XJ, XS Qi, XR Wang, JY Su (2009). Analysis of
genetic diversity of local germplasm resources for foxtail millet in Gansu. Agric
Res Arid Areas 27:129‒153
Wen QF, L Wang, WH Zhao, JW Chang (2002). Reproduction
and storage of foxtail millet germplasm in Shanxi. J Shanxi Agric Sci
30:32‒34
Xiang JS, F Xu, LX Suo, K Cheng, YC Wang, HL Meng,
JL Zhang, B Jia, DX Wang, XM Diao (2018). Comparison on the phenotypic traits
between landraces and cultivars of foxtail millet (Setaria italica (L.) P
Beauv.) in northeast China. J Plant Genet Resour 19:642‒656
Zhang AY, XM Diao, EH Guo, HP Fan, LX Wang, YH Li,
LP Cheng, YS Wu, L Zhang (2017). Research progress and major traits of foxtail
millet cultivars developed in the early-mature spring-sowing region in the past
15 tears. Sci Agric Sin 50:4496‒4506
Zhang TZ (2003). Crop Breeding. Chinese
Agriculture Press, Beijing, China