The Genetic Diversity of 69 Widely Used Chinese Sorghum Hybrids Released between the 1970s and 2010s

Abstract

Sorghum has a long history of cultivation in China. In this study, we aimed to clarify the genetic relationships and genetic variation trends in widely used Chinese sorghum hybrids which were released from the 1970s to 2010s and attempted to analyze the changes in sorghum breeding. A total of 257 alleles were detected by 51 polymorphic SSR markers among 69 widely used hybrids; an average of 5.04 alleles were detected by each marker. The average Shannon’s index and polymorphism information content (PIC) of markers were 1.39 and 0.70, respectively. Nei’s genetic diversity index continuously increased in four different breeding development stages (1973–1982, 1983–1992, 1993–2002, and 2003–2014). Genetic diversity gradually increased among the sorghum hybrids. Genetic similarity coefficients in the four breeding development stages first showed an increasing trend, and then a decreasing trend, finally stabilizing with an average value of 0.65. The genetic similarity changes in hybrids in early and late maturing areas were consistent at different breeding development stages. The genetic similarity coefficients in late maturing areas were constantly higher than those in the early maturing areas. This is related to China’s creative utilization of A₂ cytoplasmic male sterile materials in the 1990s. A cluster analysis determined that 69 hybrids were divided into two groups, A and B. Group A could be further subdivided into four subgroups. These findings could provide a reference for parental selection and hybrid breeding in sorghum improvement programs.

1. Introduction

Sorghum [Sorghum bicolor (L.) Moench] is the fifth largest major cereal crop in the world [1]. It has high levels of protein and fiber, making it a multipurpose crop widely used in the food, feed, brewing, energy, and broom industries. It is also a model crop for genomic research and functional genetic research on tropical crops [2]. Sorghum originated in sub-Saharan Africa and has since spread widely around the world. During its spread, several botanical subspecies have been differentiated, and abundant phenotypic and genetic variations have accumulated [3,4]. Sorghum is a C4 crop with high efficiency in utilizing nutrients, light, and water [5]; strong resistance to various abiotic stresses [6]; and wide cultivation in over 110 countries around the world [7]. In 2022, global sorghum production reached 60.06 million tons [8].

Chinese sorghum has a long history of cultivation, forming its own unique genetic characteristics. The collection and organization of sorghum germplasm resources in China can be traced back to the period of the 1920s–1930s, and three large-scale national germplasm resource collection and organization programs were carried out in 1956, 1978, and the 1979–1984 period, when most of the landraces were collected. Xu Guanren first introduced male sterile lines and maintainer lines (Tx3197A and Tx3197B) from the USA into China in 1956. In the early stages of heterosis research in China, restorer lines were mainly selected from Chinese landraces, and the diffusion of the varieties developed at this stage was often limited because the plants of these varieties were always too tall [9]. In 1965, Niu Tiantang, at the Sorghum Research Institute of the Shanxi Academy of Agricultural Science, pioneered the process of dwarf sorghum breeding in China by developing high-yielding dwarf sorghum hybrid varieties Jinza 5, Jinza 1, and Jinza 4. The cumulative cultivation area of these three varieties has exceeded 7.06 million hectares, promoting the development of hybrid sorghum breeding in China. In the 1970s, dwarf breeding and hybrid varieties were widely utilized in sorghum breeding, which promoted the rapid development of sorghum production in China. In this period, representative male sterile lines included Tx3197A, Ji2731A, Tx622A, 421A, A₂V4A, and 7050A, whereas representative restorer lines included Kaoliang Sanchisan, Kaffir sorghum Jinfu 1, and Hegari sorghum Ji7313, Ji7384, Xinliang 7, Jinliang 5, and Tanghui 10 [10]. A large number of sorghum hybrid varieties were released, and several large-scale seed production bases were formed. The per unit area yield of sorghum doubled from 1155 kg/ha in 1965 to 2310 kg/ha by the mid-1970s. In the 1980s, owing to the adjustment of the cropping structure in China, the planting area of sorghum was nearly halved compared to that in the 1970s. Breeding objectives changed accordingly, gradually shifting towards brewing and feed. However, during this period, the introduction of foreign germplasm resources and the application of A₂ cytoplasmic male sterile lines significantly advanced sorghum breeding in China. By the 1990s, the planting area of sorghum declined, but the total yield remained at around 5 million tons [11].

The most commonly used male sterile lines in Chinese hybrid sorghum production are A₁-type and A₂-type cytoplasmic male sterility (CMS). Commercial sorghum hybrid varieties were first successfully developed using the A₁ CMS system in the 1950s [12] and applied in sorghum breeding in China along with the introduction of Tx3197A. From 1983 to 1993, to solve the limitation of using just A₁ cytoplasm in sorghum heterosis utilization, breeders in Shanxi Province first developed the commercially available A₂-type cytoplasmic male sterile line A₂V4A [13]. A₂V4A used sorghum non-Milo cytoplasmic A₂TAM428 as a sterile cytoplasm resource and Shallu material V4 (picklet) as the recipient. Then, a series of hybrid varieties such as Jinza 12 were released. The commercial utilization of A₂ CMS in sorghum hybrid breeding solved the problem of A₂ CMS, which was unstable and could not be directly used for production, changing the long-term utilization of unique cytoplasm in hybrid sorghum breeding [14].

One of the difficulties in developing and utilizing hybrids was the lack of good restorers and maintainers. Only a few local Chinese varieties could be used as maintainers, and when considering pollen fertility and other comprehensive traits, the utilization of local varieties in hybrid breeding is challenging. After the 1970s, the use of foreign germplasm resources greatly enriched the genetic diversity of Chinese sorghum varieties, and introduced materials became the main sources of restorer lines. The CMS system used in present sorghum hybrid breeding in China mainly includes the Kafir, kaoliang, trend Kafir, and Shallu CMS-Rf genetic systems. The restorer lines mainly include trend kaoliang material with the kaoliang-type CMS-Rf genetic system. The Kafir × kaoliang mode, kaoliang × kaoliang mode, trend Kafir × trend kaoliang mode, and Shallu × trend kaoliang mode are the major heterosis utilization modes in China [9,12].

Genetic variation in crops is essential for the improvement in desirable traits, such as increased yield and tolerance/resistance to biotic and abiotic stresses. Crops with low genetic diversity are vulnerable to biotic and abiotic stresses, which can have devastating effects. Therefore, conserving genetic diversity is crucial for ensuring global food security [15]. When considering the hybrid combination, choosing parents with high genetic distance is helpful to produce superior hybrids. The continuous evaluation of genetic gains and diversity could guide breeding plans and help improve the genetic structure of sorghum. One of the main effects of breeding in the 20th century is the reduction in crop genetic diversity [16]. In past decades, a large amount of research focused on genetic diversity and genetic architecture and their effects on crop agronomic traits and resistance in sorghum [3,17,18,19]. Studies have focused on the genetic diversity of sorghum germplasms in the USA, India, and some countries in Africa; some studies focused on sorghum local varieties in China [18,19,20,21,22,23]; and some studies summarized the development of hybrid breeding in China [10,11,14]. However, more such work remains to be undertaken to explore the available genetically diverse sorghum for conservation and sustainable utilization. In our study, SSR markers were used to analyze the genetic diversity of sixty-nine widely used hybrid varieties in China. The tested hybrid varieties were released from the 1970s to 2010s. The purpose of this research was to clarify the genetic relationships and genetic variation trends in Chinese sorghum hybrids and provide a reference for parental selection and hybrid breeding in sorghum breeding.

2. Materials and Methods

2.1. Materials

Sixty-nine Chinese sorghum hybrid varieties were selected on the basis that they are or have been widely used in sorghum production in China from the 1970s to 2010s to conduct a genetic diversity analysis. Among these varieties, 19 were from Shanxi Province; 19 from Liaoning Province; 11 from Jilin Province; 10 from Inner Mongolia; 9 from Heilongjiang Province, and 1 from Hebei Province. Detailed information of the varieties is listed in

Table 1. Sorghum hybrids used in this study and their origin.

Stages	Sorghum Hybrid	Parental Combination	Release Year and Province
1 (1973–1982)	Jinza 1	Tx3197A×Jinfu 1	1973, Shanxi
	Jinza 4	Tx3197A×Jinliang 5	1973, Shanxi
	Jinza 5	Tx3197A×Sanchisan	1973, Shanxi
	Jinza 7	Tx3197A×Xin 7	1973, Shanxi
	Tongza 2	Heilong 11A×Jihui 7384	1973, Shanxi
	Jiza 52	Heilong 30A×Jihui 13	1979, Jilin
	Chiyu 7	Hei 30A×Yin 71	1980, Inner Mongolia
2 (1983–1992)	Liaoza 1	Tx622A×Jinfu 1	1983, Liaoning
	Jiza 5	Cheng 3A×7501	1983, Hebei
	Aoza 1	314A×5933	1984, Inner Mongolia
	Chiza 851	314A×7657	1987, Inner Mongolia
	Jinzhong 405	7501A×Jinliang 5	1987, Shanxi
	Jinza 11	Tx623A×Xin 7	1987, Shanxi
	Kang 4	Tx623A×Jinliang 5	1988, Shanxi
	Shenza 5	Tx622A×0-30	1988, Liaoning
	Qiaoza 2	Tx622A×L654	1988, Liaoning
	Tieza 8	Tx623A×Tieling 157	1989, Liaoning
	Shenza 6	Tx622A×5-27	1989, Liaoning
	Siza 4	2731A×140R	1990, Jilin
	Jinza 86-1	Tx623A×HM65	1991, Shanxi
	Chiza 14	Chi 12A×7663	1991, Inner Mongolia
	Aoza 2	314A×7788	1992, Inner Mongolia
3 (1993–2002)	Jinza 93	232EA/2036A×Shen 5-27	1993, Liaoning
	Jinza 12	A2V4A×1383-2	1994, Shanxi
	Huliang 1	F14A×0-30	1995, Liaoning
	Xiongza 3	Tx622A×4930	1994, Liaoning
	Jiza 76	Heilong 11A×7431-24	1995, Jilin
	Liaoza 10	7050A×LR9198	1997, Liaoning
	Siza 25	TAM428×Nan 133	1998, Jilin
	Jinza 15	Heilong 11A×Qikangqi	1998, Shanxi
	Longza 5	301A×Hahui 118	1999, Heilongjiang
	Jinza 18	7501A×R111	1999, Shanxi
	Liaoza 12	7050A×654	2001, Liaoning
	Jinza 100	7050A×9544	2001, Liaoning
	Jiza 90	4190A×9060	2001, Jilin
	Liaoza 11	7050A×148	2001, Liaoning
	Chiza 16	Fan 8A×7654	2002, Inner Mongolia
4 (2003–2014)	Liaoza 15	LA-17×LR9198	2003, Liaoning
	Jiza 97	352A×133-6-8	2004, Jilin
	Liaoza 21	363A×0-01	2005, Liaoning
	Jiza 99	TAM428A×Ji R107	2005, Jilin
	Jinza 101	F44A×363/2691	2005, Shanxi
	Longza 9	325A×Hahui 118	2006, Heilongjiang
	Jiza 118	Ji 2055A×R8063	2007, Jilin
	Liaonian 3	Liaonian A-2×R-2	2008, Liaoning
	Loangza 10	454A×Hahui 591	2008, Heilongjiang
	Loangza 11	403A×Hahui 576	2008, Heilongjiang
	Jinza 22	SX44A×SXR-30	2008, Shanxi
	Jinza 23	SX45A×SXR-30-1	2008, Shanxi
	Chiza 24	0253A×0282	2008, Inner Mongolia
	Loangza 12	Heilong 429A×Non 68	2009, Heilongjiang
	Jiza 122	Ji 2055A×R105	2009, Jilin
	Jiza 124	Ji 2055A×Ji R107	2009, Jilin
	Jinza 102	F44A×0-30 Hong	2009, Shanxi
	Jinza 103	F44A×LR233	2009, Shanxi
	Liaonian 4	LA-25×7037	2010, Liaoning
	Jinza 106	081A×580	2010, Liaoning
	Longza 13	Heilong 423A×Nong 68	2010, Heilongjiang
	Jiza 127	Ji 2055A×R117	2010, Jilin
	Chiza 28	Chi A7×7654	2010, Inner Mongolia
	Chiza 29	Chi A6×7654	2010, Inner Mongolia
	Jinza 104	Lu 45A×Z233	2011, Shanxi
	Liaonian 5	Fu A-1×Liaonian R-4	2012, Liaoning
	Liaonian 6	LA-34×0-01 Xuan	2012, Liaoning
	Suiza 7	Suibuyu 30A×Suihui 25	2012, Heilongjiang
	Chiliang 4	314A×R185	2013, Inner Mongolia
	Jinza 33	SX605A×Nan 133	2013, Shanxi
	Longza 16	Heilong 433A×Hahui 591	2014, Heilongjiang
	Suiza 8	Suibuyu 26A×Suihui 27	2014, Heilongjiang

. The planting areas of these materials were divided into “early maturing area” and “late maturing area” according to Lu’s description [24]. Materials were planted in the Dongbai experimental station of the Sorghum Research Institute, Shanxi Agricultural University in Jinzhong, Shanxi Province, China (latitude: 112.703291; longitude: 37.586570).

2.2. Genotyping

Genomic DNA was isolated from the leaves of sorghum seedlings at the 3–5 leaf stage. To avoid the impact of miscellaneous plants, leaves of 5 seedlings were mixed for DNA isolation in each entry, and DNA isolation was performed using a DNA isolation kit [Tiangen Biotech (Beijing) Co. Ltd., Beijing, China] according to the accompanying instructions. Simple sequence repeat (SSR) markers were firstly screened for their polymorphism between different materials, and finally, fifty-one markers that were evenly distributed throughout the sorghum genome and with polymorphisms in the population were selected for amplification (Table S1). Polymerase chain reaction (PCR) amplification was performed in a 20 μL mixture containing 2.5 μL 10 × Buffer (Mg²⁺), 0.3 μL dNTP (2.5 mmol·L⁻¹), 0.2 μL DNA Taq DNA polymerase (1 U·μL⁻¹), 1 μL of each primer (10 μmol·L⁻¹), 2 μL DNA (30 ng·μL⁻¹), and 13 μL ddH₂O. Amplification was carried out using a Peltier PCR system (M.J Research PTC-225) with the following thermal profile: 94 °C for 3 min; 35 cycles at 94 °C for 45 s, a suitable annealing temperature for 45 s, and 72 °C for 60 s; and 72 °C for 10 min. The PCR products were stored at 4 °C. DNA fragments were separated into 6% non-denaturing polyacrylamide gels (Acr:Bis = 29:1) at room temperature with 1×TBE buffer and visualized by silver staining.

2.3. Statistical Analysis

The genotypes of the 69 hybrid varieties were analyzed based on the results of gel electrophoresis. The presence and absence of a band at the target band position were labeled as 1 and 0, respectively. Missing data were labeled as 9. Statistical analysis was performed using NTSYSpc version 2.1a. Gene frequency was estimated with the simple matching coefficient, and the genetic similarity coefficient was estimated as $\mathrm{G}\mathrm{S}={\displaystyle \frac{\mathrm{m}}{(\mathrm{m}+\mathrm{n})}}$, with m representing the number of common bands between genotypes and n representing the number of different bands. Genetic similarity clustering was performed by the Unweighted Pair Group Method with Arithmetic Means (UPGMA), and a dendrogram was constructed. The polymorphic information content (PIC) value was estimated using Powermarker version 3.25 as $\mathrm{P}\mathrm{I}\mathrm{C}=1-\sum _{\mathrm{j}=1}^{\mathrm{i}}{\left({\mathrm{P}}_{\mathrm{i}\mathrm{j}}\right)}^2$, with P_ij representing the frequency of allele j of locus i. Effective alleles reflect the size of population genetic variation, which is equal to the reciprocal of the homozygosity of the actual population. The effective number of alleles (Ne) was estimated as $\mathrm{N}\mathrm{e}=\sum _{\mathrm{i}=1}^{\mathrm{n}}{\displaystyle \frac{1}{\sum _{\mathrm{j}=1}^{\mathrm{m}}{\mathrm{P}}_{\mathrm{i}\mathrm{j}}^2}}$, with P_ij representing the frequency of allele j of locus i, n representing the total number of detected loci, and m representing total number of alleles of locus i. Shannon’s diversity index (H) was estimated as $\mathrm{H}=-\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{P}}_{\mathrm{i}}\ln {\mathrm{P}}_{\mathrm{i}}$, with P_i representing the frequency of locus i and n representing the total number of alleles. Nei’s genetic diversity index (He) was estimated as $\mathrm{H}\mathrm{e}={\displaystyle \frac{1}{\mathrm{n}}}\sum _{\mathrm{i}=1}^{\mathrm{n}}(1-\sum _{\mathrm{j}=1}^{{\mathrm{m}}_{\mathrm{i}}}{\mathrm{P}}_{\mathrm{i}\mathrm{j}}^2)$, with P_ij representing the frequency of allele j of locus i, n representing the total number of detected alleles, and m representing the total number of alleles at locus i. Data processing was performed with MS Excel.

3. Result

3.1. Diversity of SSR Markers in Hybrid Sorghum Varieties

The use of appropriate molecular markers is important for the analysis of crop genetic diversity. In this study, a total of 257 alleles were detected in 69 hybrid varieties by 51 polymorphic markers, and the markers were distributed on all 10 chromosomes of the sorghum genome (

Table 2. Amplification of 51 pairs of SSR markers in 69 tested hybrids.

Marker	Number of Alleles (Na)	Number of Effective Alleles (Ne)	Polymorphism Information Content (PIC)	Shannon’s Genetic Diversity Index (H)
Xtxp46	5	4.13	0.76	1.46
Xtxp58	4	3.52	0.72	1.3
Xtxp75	6	4.02	0.75	1.54
Xtxp78	7	6.22	0.84	1.86
Xtxp248	5	4.11	0.76	1.47
Xtxp279	6	4.78	0.79	1.6
Xtxp302	9	6.58	0.85	1.99
Xtxp8	4	3.06	0.67	1.25
Xtxp84	5	3.64	0.73	1.44
Xtxp96	5	4.83	0.79	1.59
Xtxp100	7	6.58	0.85	1.91
Xtxp286	9	8.15	0.88	2.15
Xtxp298	9	7.76	0.87	2.12
Xtxp315	8	6.97	0.86	2.01
Xtxp31	9	7.96	0.87	2.13
Xtxp34	3	2.94	0.66	1.09
Xtxp69	6	4.46	0.78	1.62
Xtxp120	8	5.22	0.81	1.82
Xtxp266	2	1.88	0.47	0.66
Xtxp285	7	5.59	0.82	1.80
Xtxp60	3	2.51	0.60	1.00
Xtxp177	4	2.40	0.58	1.01
Xtxp212	3	1.90	0.47	0.83
Xtxp15	5	4.64	0.78	1.57
Xtxp94	4	3.04	0.67	1.23
Xtxp262	4	3.33	0.70	1.29
Xtxp17	6	5.59	0.82	1.76
Xtxp95	6	3.92	0.74	1.55
Xtxp145	4	3.45	0.71	1.29
Xtxp176	4	2.38	0.58	1.05
Xcup-37	2	1.69	0.41	0.60
Xtxp36	2	1.44	0.30	0.48
Xtxp40	4	2.43	0.59	1.11
Xtxp159	6	5.76	0.83	1.77
Xtxp168	5	4.12	0.76	1.46
Xtxp278	3	2.89	0.65	1.08
Xtxp312	5	3.23	0.69	1.38
Xtxp18	7	5.50	0.82	1.80
Xtxp47	2	1.78	0.44	0.63
Xtxp105	5	2.31	0.57	0.99
Xtxp210	4	2.65	0.62	1.16
Xtxp250	5	4.28	0.77	1.51
Xtxp354	3	2.98	0.66	1.09
Xcup-47	5	3.32	0.70	1.31
Xtxp10	6	4.13	0.76	1.55
Xtxp67	4	3.68	0.73	1.34
Xtxp287	2	1.94	0.48	0.68
Xtxp289	2	1.95	0.49	0.68
Xtxp23	8	6.18	0.84	1.90
Xtxp141	5	4.61	0.78	1.57
Xtxp217	5	4.83	0.79	1.59
Average	5.04	4.06	0.70	1.39

). The number of detected allele variations at each locus ranged from 2 to 9, with an average of 5.04. Furthermore, 80.64% of the 257 detected alleles were effective alleles. The effective allele number for each locus ranged from 1.44 to 8.15. These SSR markers could detect the diversity between different materials effectively and were suitable for genetic diversity research in this species.

The results of the genetic diversity and polymorphism information analysis are shown in Table 2. Shannon’s diversity index of 51 markers ranged from 0.48 to 2.15 with an average of 1.39; the PIC ranged from 0.30 to 0.88 with an average of 0.70. In this study, 44 markers (86.3% of the tested markers) were characterized by a high polymorphism (PIC > 0.5), whereas 7 markers (13.7% of the tested markers) showed a lower polymorphism (0.5 > PIC > 0.25). The genetic diversity between markers was different: for the marker with the highest polymorphism, Xtxp286, the number of effective alleles was 8.15 alleles, Shannon’s diversity index was 2.15, and the PIC was 0.88; for the marker with the lowest polymorphism, Xtxp36, the number of effective alleles was 1.44, Shannon’s diversity index was 0.48, and the PIC was 0.30.

3.2. Analysis of Genetic Diversity and Genetic Similarity in Hybrid Varieties Released in Different Breeding Development Stages

The genetic diversity and genetic similarity of the hybrids were estimated. The numbers of tested hybrid varieties in the four stages of sorghum heterosis utilization in China were 7, 15, 15, and 32, respectively. The average numbers of varieties released for promotion each year were 0.7, 1.5, 1.5, and 2.7. When considering Nei’s genetic diversity index in varieties released at each breeding stage (

Table 3. Alleles of SSR markers among hybrids released in different stages.

Stages	Number of Hybrids	Number of Alleles	Number of Introduced Alleles	Number of Lost Alleles *	Average Number of Alleles	Nei’s Genetic Diversity Index
1 (1973–1982)	7	232	/	/	4.55	0.66
2 (1983–1992)	15	247	20	5	4.84	0.67
3 (1993–2002)	15	242	6	11	4.75	0.68
4 (2003–2014)	32	254	12	0	4.98	0.69

* Introduced alleles and lost alleles compared to their presence in the previous stage.

), in the first stage, the total number of alleles was 232; the numbers of alleles introduced and abandoned (the genotypes used for crossing in the later stage did not contain the alleles in the genotypes used for crossing in the earlier stage) in the second stage were 20 and 5, respectively; the total number of alleles increased by 15 (from 232 to 247); and Nei’s genetic diversity slightly increased from 0.66 to 0.67. In the third breeding stage, the numbers of alleles introduced and abandoned were 6 and 11, respectively. As a consequence, the total number of alleles decreased by 5, reaching 242, while Nei’s genetic diversity index was 0.68. In the fourth stage, 12 alleles were introduced, no alleles were abandoned, and the total number of alleles reached 254, with an increase of 12 compared with the third stage. When the final stage was compared with the first stage, the total number of alleles increased by 22, and Nei’s genetic diversity index reached 0.69.

The average numbers of alleles in the four different breeding stages were 4.55, 4.84, 4.75, and 4.98, respectively (Table 3). As already reported, there was an upward trend in the average number of alleles from the first stage to the second stage, a slight decrease in the third stage, and a rebound in the fourth stage. Nei’s genetic diversity index had been slightly increasing in all different breeding stages. Although some alleles were abandoned at different breeding stages, the overall number of newly introduced alleles was greater than the abandoned alleles, and the diversity level of sorghum varieties increased gradually.

When comparing the genetic similarity coefficient of the varieties (Table S2), the mean genetic similarity coefficient was 0.66. The genetic similarity between Jinza 22 and Jinza 102 was the highest with a genetic similarity coefficient of 0.83. Meanwhile, the genetic similarity between Jinza 104 and Longza 9 was the lowest with a genetic similarity coefficient of 0.52. The mean genetic similarity coefficients between hybrids in the four breeding stages were 0.65, 0.68, 0.66, and 0.66, respectively (

Table 4. Genetic similarity coefficient between hybrids released in different stages.

	First Stage (1973–1982)	Second Stage (1983–1992)	Third Stage (1993–2002)	Fourth Stage (2003–2014)
Nationwide	0.65	0.68	0.66	0.66
Early maturing area	0.68	0.70	0.64	0.66
Late maturing area	0.71	0.72	0.67	0.68
Shanxi Province	0.71	0.76	0.63	0.69
Liaoning Province		0.76	0.69	0.70

). The genetic similarity coefficient first showed an increasing trend, and then a decreasing trend, and remained stable at last, reaching the highest value in the second breeding stage. The genetic similarity changes in the hybrid varieties in different breeding stages in early and late maturity regions were consistent, i.e., they all showed an upward trend upward, and then a downward trend, and finally an upward trend. The genetic similarity coefficient in late maturing areas was higher than that in the early maturing areas at all stages; this was probably related to the creative development of A₂ cytoplasmic hybrid varieties for sorghum production in the 1990s in China.

3.3. Cluster Analysis of Hybrid Sorghum Varieties

After the estimation of the genetic diversity and similarity of the hybrids, the relationships among the hybrids were further analyzed. Genetic similarity clustering was performed using the UPGMA method, and a dendrogram was constructed. We determined 0.62 as the threshold for genetic similarity coefficient according to the clustering pattern. At this threshold, different materials can be distinguished, albeit without too many clusters. Finally, 69 sorghum hybrid varieties were divided into two groups, A and B (Figure 1;

Table 5. The clustering results of the 69 sorghum hybrids.

Group	Sorghum Hybrids	Number of Varieties
A	Subgroup I: Chiza 851, Aoza 1, Aoza 2, Chiliang 4, Chiza 16, Chiza 14, Chiza 28, Chiza 29, Jiza 118, Jiza 124, Jiza 122, Jiza 127, Chiza 24, Jinza 101, Jinza 103, Huliang 1, Jinza 12, Jinza 22, Jinza 102	19
	Subgroup II: Liaonian 6, Liaonian 4, Liaonian 3, Liaonian 5, Liaoza 1, Shenza 5, Shenza 6, Xiongza 3, Qiaoza 2, Jiza 97, Jiza 90, Siza 4, Suiza 7, Suiza 8	14
	Subgroup III: Jiza 5, Jinza 4, Kang 4, Jinza 5, Jinza 18, Jinzhong 405, Jinza 86-1, Tieza 8, Jinza 7, Jinza 11, Jinza 1, Jinza 104, Jinza 23	13
	Subgroup IV: Tongza 2, Jinza 15, Jiza 76, Jiza 99, Siza 25, Jinza 33, Liaoza 21, Liaoza 10, Liaoza 11, Jinza 100, Liaoza 12, Jinza 93, Liaoza 15, Jinza 106	14
B	Chiyu 7, Jiza 52, Longza 16, Longza 5, Longza 9, Longza 12, Longza 13, Longza 10, Longza 11	9

). In group A, 60 varieties were present, and this group could be further divided into four subgroups. Group B included nine hybrid varieties, all from early maturing areas, including one from Inner Mongolia, one from Jilin, and seven from Heilongjiang. All of the parents of these hybrid varieties were derivatives of Kubanskoe krasnoe (1677б or 48).

Subgroup I of group A contained 19 hybrid varieties. Among them, nine varieties were released from the early maturing area, Inner Mongolia. The female parent of these materials was 314A or its derived sterile lines. Four varieties were from Jilin Province, with the common female parent being Ji2055A, which was transformed from the offspring of 314B and Shallu sorghum-derived material 871300B. Five and one varieties were from the late maturing areas, the Shanxi and Liaoning provinces, respectively. These six varieties were all derivatives of Shallu sorghum with A₂ CMS.

Subgroup II of group A contained 14 varieties. Three and two varieties were from the early maturing areas of Jilin and Heilongjiang provinces, respectively. Nine varieties were from the late maturing area of Liaoning Province. Female parents of both Jiza 90 and Jiza 97 were derivatives of TAM428. Among the varieties from Liaoning Province, five varieties were developed by Tx622 in the period of the 1980s to 1990s, and the other four varieties were varieties with floury endosperm.

Subgroup III of group A contained 13 varieties, all from late maturing areas. Among these varieties, 11 were from Shanxi Province, 1 was from Liaoning Province, and 1 was from Hebei Province. The female parents of these varieties included Tx3197A, Tx623A, 7501A, 45A, and Cheng3A; the male parents of these materials included Jinfu 1 and Jinliang 5 and their derivatives.

Subgroup IV of group A included 14 hybrid varieties. Among them, three and eight were from the late maturing areas of Shanxi Province and Liaoning Province, respectively, and three were from the early maturing area Jilin Province. Tongza 2, Jinza 15, and Jiza 76 had the common female parent Heilong 11A; Jiza 99, Siza 25, and Jinza 33 had the common male parent Nan 133. Among the other eight varieties from Liaoning Province, Liaoza 10, Liaoza 11, Liaoza 12, and Jinza 100 were developed using the common sterile line 7050A; the restorer line of both Liaoza 15 and Liaoza 10 was LR9198. The restorer line of Liaoza 21, the 0–10 line, was an improved line from the 5–27 line, the restorer line of Jinza 93. The parents of Jinza106 were improved from the parents of Jinza 100.

In the early maturing area, as already reported, the hybrid varieties from Inner Mongolia were clustered in subgroup I of group A; the varieties from Jilin Province were clustered in the I, II, and IV subgroups of group A, and the varieties from Heilongjiang Province were mainly clustered in group B. In the late maturing areas, the varieties from Shanxi Province were clustered in subgroups I and III of group A, and the varieties from Liaoning Province were clustered in subgroups II and IV of group A. The sorghum hybrids from the early and late maturing areas could not be completely separated because of the mutual introduction of parental lines in different ecological areas.

4. Discussion

4.1. SSR Marker Performed High Polymorphism in Hybrid Variety Population

Along with the development of molecular genetics, the application of molecular markers such as AFLP, RFLP, SSR, DArT, and SNP has promoted the development of QTL mapping, a genetic diversity analysis, and variety fingerprinting in sorghum [2]. SSR markers have been widely used in molecular genetic research for their wide distribution, easy detection, high polymorphism, and co-dominant feature [25,26,27,28,29].

In this study, SSR markers were used to analyze the genetic diversity of sorghum hybrid varieties released in China from the 1970s to 2010s. In previous research, Smith [20] detected 956 alleles using 167 SSR markers in 63 sorghum hybrid varieties that were widely grown in the USA. The number of alleles detected by SSR markers ranged from 4.44 to 4.57. The marker polymorphism in Smith’s study was similar to that found in our research, and both studies used cultivated hybrids as experimental materials. Dickson [18] used 10 SSR markers to detect 27 sorghum accessions collected from agroecological regions I and II in Zambia. The SSR markers generated 2 to 9 alleles per locus and a total of 44 alleles, with an average of 4.4 alleles per marker. The total number of alleles was less than what we found because the materials and markers they used were both less than ours; nonetheless, the number of alleles detected by single markers was similar to ours. Folkertsma [21] detected 123 alleles in 100 guinea-race subspecies of sorghum materials, which were predominantly inbred and collected from 10 African countries; they detected an average of 5.86 polymorphic loci per marker. Menz [30] detected the genetic diversity of 50 inbred lines which were widely used as hybrid parents in the USA in the past 50 years with seven SSR markers, and the detected number of alleles ranged from 2 to 19, with the average of polymorphic alleles being 7.8 per marker, which is higher than that in other research. In Mamo’s study [22], an average of 7.2 alleles per marker were detected in a population consisting of 100 accessions from the Amhara; Dire Dawa; Oromia; Southern Nations, and Nationalities, and Peoples and Tigray regional states of Ethiopia. In both our research and others’, SSR markers had a high polymorphism in sorghum germplasm, and this was helpful for us to explore the genetic diversity and excellent alleles of the germplasm.

Polymorphic alleles were also detected by Permual [31] using five sub-species and nine intermediate species of sorghum as research materials, and they detected an average of 13.86 alleles per marker. Ghebru [32] detected 208 alleles in 28 Eritrean landraces and 32 world sorghum germplasm accessions using 15 SSR markers, with an average of 13.87 alleles per marker, and the number of alleles detected in these two studies was much higher than the research described previously; this might be because of the extensive sources of the research materials. Using 6977 SNP markers, Mudaki [33] studied the genetic diversity of 169 sorghum materials from Kenya and Tanzania, and the PIC of the markers was 0.31. In a population of 184 inbred lines from Ethiopia, Mindaye [34] detected an average PIC of 0.25 with Genotyping By Sequencing (GBS) markers. The number of alleles detected by SNP markers in these two studies was lower than that detected by SSR markers, but the marker density of SNP markers was much higher than that of SSR markers. The genetic background of the experimental materials and the type of markers could affect the results of the genetic diversity analysis. The polymorphism of SSR markers performed better in several different materials and was enough to distinguish among different accessions, proving their value in germplasm research.

4.2. The Overall Genetic Diversity of Sorghum Varieties in China Was Gradually Increasing

Until now, little work has been reported for the genetic diversity of Chinese hybrids. In the present study, during four breeding stages from the 1970s to 2010s, the genetic diversity among sorghum breeding materials in China increased gradually. It is worth mentioning that in our research, the numbers of hybrid varieties included in the genetic diversity analysis in the four breeding stages were 7, 15, 15, and 32, and the numbers of their parental genotypes were 11, 21, 26, and 56, respectively. Since the population size was the mediating factor in many hypotheses aimed at explaining global patterns of genetic diversity [35], the difference in the numbers of hybrid varieties and their parents might be one of the reasons for the increase in genetic diversity. But this difference could not be simply attributed to the differences caused by biased sampling, and the difference in the samples was not the only reason for the change in the allele number; we thought the genetic diversity of the materials should be more important. In the early stage, only a few germplasms were used as parents, which limited the genetic diversity of hybrids. With the development of breeding and innovation of germplasms, more genotypes were involved, which led to the increasing genetic diversity in the hybrids. In the other way, the increase in the parent material numbers did not correspond to the degree of genetic diversity change, indicating a certain degree of homogeneity between parents [36]. The innovation of breeding parental materials is lagged compared to variety selection work. Based on the studies of wheat and other field crops, van de Wouw [37] indicated that the genetic diversity of varieties developed in the 1960s decreased by 6% with respect to the 1950s, while in the 1970s, due to the genetic bases of varieties, the genetic diversity of the released varieties increased. Along with van de Wouv’s results, the fact that the genetic diversity of varieties has not significantly decreased in a long-term scale, and concerns about the disappearance of the world’s varietal wealth of crop plants and the establishment of a worldwide network of international gene banks prevent the continuous decrease in the genetic diversity of crop plants. Based on the genetic diversity analysis of USA sorghum hybrids, Smith [20] indicated that the diversity level of actual USA varieties slightly decreased, but it was not a continuous decrease, as it was characterized over time by an increasing trend followed by a decreasing trend. It should be emphasized that increasing, decreasing, and maintaining genetic diversity levels did not just a simple increase or decrease in alleles, but each case was accompanied by the introduction of new alleles and the elimination of old alleles [20,37]. The genetic diversity of hybrid varieties was influenced by germplasms, breeding plans, the demand for disease, insect and weed resistance, market demand, and quality control.

Based on the major sterile lines used, the released time of hybrid varieties used in this study could be divided into four stages. Tx3197A (1973–1982) and Tx6A (1983–1992) were the dominant sterile lines in the first and second stages; in the third stage (1993–2002), the utilization of the A₂ cytoplasm began to diffuse; and in the fourth stage (2003–2014), the selected sterile lines were utilized more independently in hybrid sorghum breeding. In the first stage, the breeding and promotion of hybrid sorghum varieties were carried out extensively, and landraces belonging to the Kaoliang subspecies were widely used in crosses for hybridization with introduced materials such as Kaffir and Hegari [9]. Significant genetic differences and small similarity coefficients were present among the varieties released at this stage. During the second stage, the extensive planting of wheat and rice improved people’s living standards, and sorghum gradually changed from a major food crop to a crop used as feed and for processing raw materials; additionally, the planting area of sorghum in China declined, and the scarcity of suitable restorer lines led to difficulty in applying CMS systems other than A₁ cytoplasm [38], restricting the diversity of genetic resources, especially the maintainer lines used in breeding programs [30,39]. Only 9% of Chinese landraces could be used to find restorer lines, and approximately 14% of the restorer lines were restorers for the A₁ cytoplasm [9]. Although the Tx6A system was introduced during the second stage, the restorer lines were genetically unique and homogeneous, and many of these materials were widely and repeatedly used in hybrid combination pairing. At this stage, the genetic similarity coefficient among hybrids increased, and genetic diversity decreased. Comparing the breeding procedure of Chinese hybrid sorghum with that in the USA, in the period of the 1980s–1990s, the genetic diversity of hybrid varieties in the USA increased slightly because of the new alleles introduced in the 1990s, which was the peak activity period of the Sorghum Conversion Program for sorghum variety improvement [40,41,42,43]; in this period, a large number of elite alleles were introduced and preserved in subsequent breeding projects [20]. The introduction of new germplasms had a significant effect on improving the genetic diversity of hybrid varieties in the USA.

During the third stage, foreign sorghum germplasms were extensively introduced to China, improving the diversification of hybrid varieties [9]. Breakthroughs have been made in the utilization of A₂ cytoplasmic male sterile lines, which improved the genetic diversity of germplasms and hybrid varieties [14]; as a consequence, similarity among hybrids decreased. In the fourth stage, although a large amount of hybrid varieties developed, no breakthrough progress has been made in the improvement in restorer and sterile lines. The main focus of agricultural research at that time was staple crops like rice, wheat, and corn, and limited attention was paid to other crops. The genetic diversity of hybrids was basically consistent with the previous stage. A similar trend in genetic diversity in USA sorghum hybrids was seen from the 1990s to 2000s. At this period, the major objective of sorghum breeders in the USA changed from introducing new genes to screening introduced genes; at the same time, the activity of the Sorghum Conversion Program dominated by the US Department of Agriculture and Texas A&M University decreased [20]. The slowdown of the introduction of new germplasms eventually influenced the change in genetic diversity of sorghum hybrids in both the USA and China. The change in the genetic diversity of hybrid varieties was the ultimate result of germplasms being introduced and the selection of breeders and genetic drift; the goals of all of these activities were increasing the hybrid yield and completing the breeding objectives. Due to the limitations in research materials, we only studied varieties developed between the 1970s and 2010s; since more hybrid varieties have been developed in recent years, a continuous evaluation of the genetic diversity of these material is very meaningful for breeding research.

4.3. Hybrid Varieties Could Be Divided into Two Major Groups with No Significant Correlation between Group Differentiation and Regional Origin of Varieties

To distinguish the genetic similarity of the hybrid sorghum varieties, a cluster analysis was performed, and widely diffused hybrid varieties of sorghum in China from the 1970s to 2010s were divided into two groups. The relationship between these hybrids and their parents was discussed. The differentiation of hybrid varieties was not significantly correlated to their geographical origin, and varieties from the early maturing area and late maturing area could not be completely separated. This result was due to the crosses among different ecologically adapted parents in the breeding procedures of Chinese sorghum. Consistent with our result, when estimating the genetic diversity of sorghum lines in a USA backcross-nested association mapping population, Corzier [44] indicated that the elite lines were always grouped based on their heterotic groups, seeds, or pollinators because of selection, recombination, and admixture over many generations in hybrid breeding programs. Based on the genetic diversity analysis of 160 sorghum lines, Silva [45] drew conclusions similar to ours, i.e., the genetic structure of each material was highly consistent with its own pedigree relationship. In Smith’s research [20], the clustering result also showed a clear relationship between the pedigree and breeding period of the hybrid varieties. Clustering was also affected by the breeding companies which released the hybrids, a situation different from the one in China. This aspect was mainly due to the fact that the development of commercial hybrid varieties in the USA was mainly carried out by seed companies, and the technical protection between companies led to certain differences in the germplasms used, while a large part of varieties in China were developed by public research institutions and universities, with a level of germplasm exchange higher than that in the USA. Zhang’s research [46] suggested that genetic variation in Chinese sorghum germplasm resources mainly existed among materials within a region or within ecological zones. In Mudaki’s study [33], 169 sorghum accessions were divided into 10 subgroups and formed six to seven main branches; the division of subgroups was based on regional characteristics, and there were significant genetic differences among subgroups. Adugna [47] analyzed 160 Ethiopian germplasm resources and found that the genetic structure of each population was influenced by human domestication, migration, and climate change. The research materials used in the studies reported by Zhang, Mudaki, and Adugna were local varieties, and there was no obvious genetic relationship between these materials. An obvious correlation existed between genetic distance and geographical distance. In modern breeding programs, these germplasm resources were collected, organized, re-hybridized, and selected. These human activities broke the isolation of genetic resources in the original region and promoted the recombination of genetic resources.

5. Conclusions

This study used SSR markers to analyze the genetic diversity of 69 sorghum hybrid varieties selected for the main sorghum-producing areas in China from the 1970s to 2010s. The widely diffused hybrid varieties in production reflect the breeding trend to a certain extent. The study of genetic diversity at the molecular level, the identification of germplasms, and the division of genetic relationships could be used to guide the genetic breeding of sorghum hybrids. The main conclusions are as follows:

An analysis of the genetic diversity of widely diffused hybrid varieties of sorghum in China from the 1970s to 2010s was conducted, and the genetic relationships and genetic variation trends in sorghum hybrids in China were assessed; the results indicate that new germplasms and genetic variations play important roles in the breakthrough progress of breeding. In sorghum breeding programs, the continuous introduction and evaluation of new genetic variations are essential for long-term improvement. The continuous evaluation of the genetic diversity of sorghum hybrids and their parents is beneficial for planning breeding projects and the sustainable development of sorghum breeding. The results of this study could provide a reference for further parental paring and hybrid breeding directions.