The Breeding of Waxy Sorghum Using Traditional Three-Line Method and Marker-Assisted Selection

Abstract

Sorghum (Sorghum bicolor) exhibits drought resistance and environmental adaptability, making it a crucial cereal crop for semi-arid regions. It has a wide range of uses, including as food, feed, brooms, alcohol production, and bioethanol. In particular, Taiwan imports nearly 50,000 tons of sorghum annually, primarily for the production of sorghum liquor. However, the ideal raw material for high-quality sorghum liquor is waxy sorghum, and not all sorghum varieties imported or promoted in Taiwan are of this waxy type. Consequently, there is a shortage of sufficient waxy sorghum raw materials to meet the demands of the Taiwan market. The occurrence of waxy sorghum (wx) is caused by the mutation of granule-bound starch synthase I (GBBS I), and there are currently several known types of mutants, carrying different wx^a, wx^b, and wx^c waxy alleles. Among them, wx^c is a novel mutation type, and in native sorghum in Taiwan, individuals with the waxy allele wx^c have been found. The three-line method is a commonly used breeding strategy, which simplifies the process of emasculation to obtain hybrid F₁ offspring. In this study, imported sorghum variety Liangnuo No.1 (with male sterility), native glutinous sorghum variety SB6 from Taiwan (carrying the wx^c waxy allele), and sorghum reference genome variety BTx623 were used as research materials. The goal was to use the three-line method to produce waxy sorghums, including the male sterile line (A-line), male sterile maintenance line (B-line), and fertility-restoring line (R-line). The breeding results showed that by using backcross breeding, molecular-assisted selection, and traditional field selection methods, high-quality three-line materials (A-, B-, R-lines, named CNA1, CNB1 CNR1, respectively) and F₁ hybrid (CNH1) with favorable agronomic traits and yield quality were successfully obtained.

1. Introduction

Sorghum is currently one of the world’s top five cereal crops, ranking just after maize, rice, wheat, and barley [1]. In 2021, sorghum was primarily produced in countries such as the USA (11,374,900 tons), Nigeria (6,725,000 tons), India (4,810,000 tons), Ethiopia (4,450,000 tons), Mexico (4,370,064 tons), Sudan (3,530,000 tons), Argentina (3,319,341 tons), and China (3,000,000 tons) [2]. Apart from being a staple food, sorghum also serves as a crucial raw material for feed, fiber, bioethanol, and alcohol production. Sorghum belongs to the Poaceae family, and it is an annual herbaceous plant, which thrives in warm climates. Its origin is traced back to approximately 5000 years ago in east Africa [3]. With its long history of cultivation and adaptation to various climates, sorghum can grow well in tropical and temperate regions. It exhibits broad adaptability to environmental conditions, thriving in high-temperature and arid environments.

Being a tropical plant, sorghum typically requires higher temperatures for optimal growth. It stands out among the Poaceae crops as one of the most resilient to extended periods of drought and poor soil quality [4]. Consequently, it is a crucial crop for alleviating food shortages, especially in semi-arid regions. Sorghum is known for its ease of cultivation, drought resistance, wide adaptability, low water requirements, and short growth cycle. It has lower cultivation costs compared to other crops. As a result, in addition to being a staple food in Africa and India, sorghum production in USA is often geared toward energy production and animal feed. Sorghum used as an energy crop is utilized for bioethanol production, addressing energy crises [5]. Moreover, countries such as Argentina, Mexico, and Australia primarily produce sorghum for animal feed. Sorghum is an excellent model crop for studying plant resilience to drought and other adverse conditions. It is also considered the best drought-resistant C4 model crop due to its relatively small genome of about 7.3 million base pairs (Mb) [6].

Sorghum is widely used in food processing, especially as a staple and secondary food source in semi-arid countries. It is utilized in various food products, such as bread (in India, Mexico, Ethiopia, Sudan), porridge (in Africa and India), steamed dishes (in west Africa and Asia), boiled dishes (in Mali), and snack foods (in India), among others [7]. Sweet sorghum primarily emphasizes stalk yield over grain yield because it can produce high-sugar juice from its stalks. Sweet sorghum grows rapidly, has a short growth period, and accumulates sugar quickly. The juice extracted from sweet sorghum stalks is rich in carbohydrates, such as sucrose, glucose, fructose, maltose, etc., making it suitable for conversion into ethanol as a biofuel [8]. Conversely, waxy sorghum places its primary emphasis on grain harvesting. Although sorghum grains contain starch, the conversion of starch into ethanol for use as a biofuel is less economically beneficial [9]. In Taiwan, sorghum is primarily imported and mainly used for animal feed and alcoholic beverage production. Sorghum holds a significant position in the alcohol production industry in Taiwan, with the value of Taiwan sorghum liquor production increasing year by year. Additionally, many indigenous alcoholic beverages in Africa and Latin America are also made from sorghum.

In terms of the main effective gene (Waxy gene, Wx) controlling straight-chain starch in sorghum, the one encoding granule-bound starch synthase I (GBSSI) is known to have three alleles [10,11,12]. The structure of non-waxy alleles in sorghum is highly similar to non-waxy alleles in other cereal species, such as rice (Oryza sativa), wheat (Triticum aestivum), barley (Hordeum vulgare), foxtail millet (Setaria italica), and maize (Zea mays). These alleles typically contain 14 exons. Intron 1 is usually small and of similar size among different species. However, non-waxy alleles in each species exhibit significant differences in other aspects [13,14].

In rice, waxy mutations are caused by point mutations within intron 1, leading to reduced levels of GBSSI mRNA and protein [15,16]. In wheat, non-waxy alleles in all three genomes encompass various sizes and positions of deletions in the waxy gene, the presence of transposons [17], and SNPs resulting in inactive GBSSI [18]. In barley, the waxy gene mRNA levels are significantly reduced due to deletions in the promoter and 5′ untranslated region. There is a 403 bp deletion in exon 1 and a 193 bp insertion in intron 1 [19]. In foxtail millet [20] and maize [21,22], the gene expression is reduced due to transposon insertions.

In sorghum, two waxy mutations have been identified, examining the presence or absence of the GBSS I protein (Sb10g002140), known as wx^a and wx^b [11]. The wx^a allele has a large sequence insertion in the third exon, which results in the loss of the GBSS protein, having a consistent effect. On the other hand, the wx^b allele is a missense mutation, which converts glutamine to histidine, altering the functional structural region of the starch synthase enzyme [10]. Sattler et al. [12] noted that wx^a and wx^b can be distinguished using molecular markers. Kawahigashi et al. [13] conducted a study on 337 sorghum accessions and found 17, which tested negative for iodine staining, confirming them as waxy genotypes, with 16 being identified as wx^a.

A novel waxy allele, wx^c, was discovered in local sorghum varieties in Taiwan. This is a new waxy allele, having a point mutation within the exon–intron boundary splice site, specifically a G to C nucleotide change at the 5′ splice site of intron 10–exon 11 boundary [13]. This mutation is likely to reduce the expression of the waxy gene. Functional DNA markers for wx^c can distinguish it from other alleles, allowing for the identification of heterozygous non-waxy and waxy starch granule-bound starch synthase I genotypes. Currently, the wild-type wx gene is known, along with three recessive waxy alleles, wx^a, wx^b, and wx^c. Most waxy sorghum varieties carry wx^a or wx^b alleles. wx^a does not produce the GBSS I protein, but wx^b can produce the GBSS I protein; however, both are unable to synthesize amylose effectively. The wx^c allele has been found exclusively in local sorghum varieties in Taiwan [11,13].

The three-line breeding system, comprising the A-line sterile line, the B-line maintainer line, and the R-line restorer line, is a crucial strategy in modern crop breeding [23]. One of its paramount advantages lies in its application within hybrid systems, alleviating the complexities associated with emasculation and thereby saving time and resources. In the context of the three-line system, the A-line sterile line typically carries a specific genotype, rendering it incapable of producing pollen, thus lacking reproductive ability through self-pollination. Consequently, breeders are relieved of the laborious task of manual emasculation, ensuring the purity of hybrid offspring. Simultaneously, the B-line maintainer line plays a pivotal role in preserving the sterility of the A-line, shielding it from contamination by foreign pollen during cultivation. Lastly, the presence of the R-line restorer line enables the restoration of fertility in the progeny of hybrids, facilitating hybrid breeding [24].

The primary objectives of this research involve utilizing backcross breeding, marker-assisted selection, and traditional field selection methods to create a set of waxy sorghum materials through the application of the three-line approach. These sorghum materials, including the A-, B-, R-lines, and hybrid F₁, are expected to show favorable agronomic traits and yields. The underlying hypothesis suggests that the integration of the three-line method with molecular-assisted selection and traditional field selection can produce suitable materials for the purpose of breeding waxy sorghum.

2. Materials and Methods

2.1. Breeding Materials

2.1.1. Liangnuo No.1 F₄ Inbreeding Line

The inbreeding lines (F₄) of Liangnuo No.1 (LN1) were developed through bagging self-pollination of the hybrid LN1. The F₂ population was subjected to selection based on the pedigree method for subsequent agricultural and morphological traits. Selection continued in F₂, F₃, and F₄ generations, resulting in the development of 29 stable recombinant inbred lines; some of them were sterile.

2.1.2. Taiwanese Native Waxy Sorghum SB6

Taiwanese Native Waxy Sorghum SB6 was collected from Taiwo Township, Kaohsiung County. This native sorghum variety, originally used by indigenous people for brewing, was selectively bred through four generations of continuous self-pollination to develop the waxy line SB6, with a plant height of nearly 200 cm.

2.1.3. BTx623

The dwarf variety BTx623, used for the complete sorghum chromosome genome sequencing, has a plant height of approximately 110–120 cm, is less sensitive to day length, and flowers in 120 days.

2.1.4. F₅ Waxy Lines from the TCR170617/SB6

TCR170617 is a restorer line, which can help restore fertility. TCR170617 was hybridized with waxy sorghum SB6 before, resulting in F₅ lines after four generations of self-pollination. These F₅ lines carry both the restorer gene and waxy gene.

2.1.5. Crop Management

The experiment took place at the Chiayi branch station of the Taiwan Agricultural Research Institute (located at 23°29′04″ N, 120°28′04″ E). The soil in this area is characterized as silty loam. The entire breeding process extended from 2015 to 2022, with the first cropping season typically commencing in March and concluding in June, featuring temperatures ranging between 20 and 30 °C. The second cropping season typically began in September and ended in December, with temperatures ranging from 18 to 28 °C. All experiments were conducted in compliance with the standard sorghum recommendations in Taiwan.

2.2. DNA Extraction and Primer Sequence

2.2.1. DNA Extraction and PCR Conditions

The DNA extraction protocol followed Wu et al. [25]. PCR was performed using a thermocycler (GeneAmp PCR System 9700, PerkinElmer Corp., Norwalk, CT, USA), as per the following conditions: initial denaturation at 94 °C for 5 min; 35 cycles of amplification at 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 2 min; and a final extension at 72 °C for 5 min. A 1× TAE buffer and a 2.5% Super Fine Agarose (SFR) gel were prepared for gel electrophoresis. The voltage was set at 250 V, and the process ran for 15 min.

2.2.2. Primer Sequence for Foreground Selection

The primer sequences of wx^a, and wx^c were designed based on Ref [13]; the primer combinations were used for amplification. Sequencing of the wx allele was conducted for both LN1 and SB6 using ABI 3730XL Genetic Analyzers (Thermo Fisher Scientific, Waltham, MA, USA). The primer sequences used are as follows:

wx^a allele

Primer sequence:

wx-F: 5′-GGCCTGGATTCAATGTTCTT-3′

wx-R: 5′-GCAGCTGGTTGTCCTTGTAG-3′

wx^a-F: 5′-CGTGGCGAGATCAAACTCTA-3′

wx^c allele

Primer sequence:

wx^c-F: 5′-GCTGGTTCTGAGTGCAACAA-3′

wx^cWT-R: 5′-ACTTCTTCTTGCCAGTGACC-3′

wx^c-R: 5′-ACTTCTTCTTGCCAGTGACG-3′

2.3. Yield Trials

Primary yield trials were conducted using the A-(CNA1), B-(CNB1), and R-(CNR1) lines, while advanced yield trials included the addition of F₁ hybrids (CNH1).

2.3.1. Preliminary Yield Trials

In the first cropping season of 2022, preliminary yield trials were conducted for the A-, B-, and R-lines. The trials were designed using a randomized complete block design (RCBD) with two replicates. Each line was planted with 100 plants, organized in 2 rows per line, with 50 plants per row, and a row spacing of 30 cm × 15 cm. At maturity, 20 plants from each line were harvested separately to assess the agronomic characteristics and grain yield performance.

2.3.2. Advanced Yield Trials

In the second cropping season of 2022, advanced yield trials were conducted for the A-, B-, and R-lines and F₁ hybrids. The trials were designed using a randomized complete block design (RCBD) with four replicates. Each line was planted with 100 plants, organized in 2 rows per line, with 50 plants per row, and a row spacing of 30 cm × 15 cm. At maturity, 20 plants from each line were harvested separately in each plot to assess the agronomic characteristics and grain performance within each plot.

2.3.3. Statistical Analysis

Data were analyzed using the R software (version 4.3.1). Analysis of variance (ANOVA) was performed, and the least significant difference (LSD) method was used for multiple comparisons.

3. Results

3.1. The Waxy Alleles wx^a and wx^c

In order to utilize molecular markers in the process of marker-assisted backcross breeding for waxy sorghum, a molecular marker analysis and testing of the alleles on the waxy gene (wx) were first conducted on the relevant parental materials. A sequencing analysis of the alleles on the waxy gene (wx) was also performed to confirm the mutation sites and nucleotide sequences.

For the analysis of the wx^a allele, the primers wx-F, wx-R, and wx^a-F were used to amplify DNA molecular fragments from the chromosomes of waxy sorghum varieties SB6 and LN1, as well as non-waxy sorghum varieties TCR2798 and TCR170617 (as control). Among these, SB6, TCR2798, and TCR170617 amplified a 523 bp molecular fragment, while LN1 simultaneously amplified two fragments of 523 bp and 615 bp (Figure 1). This indicates that SB6, TCR2798, and TCR170617 do not carry the wx^a allele, while LN1 is a heterozygote.

In the analysis of the wx^c allele, two sets of primers, wx^c-F and wx^cWT-R, as well as wx^c-F and wx^c-R, were used to analyze SB6, LN1, TCR2798, and TCR170617. Under the amplification of wx^c-F and wx^cWT-R, LN1, TCR2798, and TCR170617 produced approximately 1500 bp molecular fragments (Figure 2A). However, under the amplification of wx^c-F and wx^c-R, only SB6 and LN1 were successfully amplified to produce approximately 1500 bp fragments (Figure 2B).

Combining the results obtained from both sets of wx^c primers, it is evident that SB6 is a wx^c homozygote, LN1 is a heterozygote, while TCR2798 and TCR170617 do not possess the wx^c allele. This also indicates that LN1 is a heterozygous genotype with both wx^a and wx^c alleles. SB6 carries two alleles, both of which have wx^c, while TCR2798 and TCR170617 do not have any waxy alleles (wx^a and wx^c).

Additionally, nine sets of primers were designed and used to amplify the entire GBSSI gene segment of TCR2798, TCR170617, SB1, SB6, and LN1. SB1 is another native Taiwanese waxy sorghum accession, which also contains the wx^c allele. The results indicated that the sequences of TCR2798 and TCR170617 were consistent with the wild type (non-waxy sorghum). However, SB1 and SB6 exhibited a single-point mutation (G to C) at the wx^c mutation site (1563 bp), confirming that they both carried the wx^c allele. Furthermore, LN1 only showed a single-point mutation at the wx^c mutation site, indicating the presence of a heterozygous wx^c allele (Figure 3).

3.2. The Selection of A-Line

The goal of A-line selection was to produce a sterile line with a background very similar to BTx623 and possessing the wx^c waxy allele (Figure 4). In the first cropping season in 2015, LN1 was used as the female parent and BTx623 as the male parent for artificial hybridization through emasculation and pollination. This led to the creation of 121 hybrid F₁ seeds. In the second cropping season in 2015, 24 F₁ plants were grown as female parents and BTx623 as male parents for artificial backcrossing. This resulted in 280 BC₁F₁ seeds.

In the first cropping season in 2016, 222 BC₁F₁ plants were obtained, and during their growth, those with favorable growth characteristics and similarity to BTx623 were selected for the wx^c allele. Out of these, 42 plants exhibited heterozygous combinations of the wx^c allele (Wx^c/wx^c) and were subjected to artificial backcrossing with BTx623. This produced 200 BC₂F₁ backcross seeds. BC₂F₁ plants were grown, and in the second cropping season in 2016, 80 plants showing heterozygous combinations of the wx^c allele (Wx^c/wx^c) and having favorable growth characteristics similar to BTx623 were selected for further artificial backcrossing.

The result was the generation of BC₃F₁ backcross seeds, which were planted in the first cropping season in 2017. The BC₃F₁ plants underwent wx^c allele selection, leading to the selection of 96 plants exhibiting heterozygous combinations of the wx^c allele (Wx^c/wx^c). These 96 waxy plants were transplanted into the field and bagged during flowering, and they were pollinated with pollen from BC₃F₂ (B-line) male parents. In the second cropping season in 2017, approximately 100 BC₃F₂ plants were subjected to wx^c allele selection. Among these, 41 plants exhibited homozygosity of the wx^c allele (wx^c/wx^c). These 41 waxy plants were transplanted into the field and bagged during flowering. At the same time, they were pollinated with pollen from the male parents of the male sterile maintenance line BC₃F₂ (B-line), resulting in BC₃F₃. Starting in 2018, the 20 BC₃F₃ plants were mixed to form a male sterile population. Inferior single plants were eliminated, and three to five plants with characteristics similar to BTx623 were selected. Additionally, the male sterile maintenance line was used as the male parent for pollination to propagate the line. In 2021, the selected superior male sterile line was designated as CNA1 (Figure 5).

3.3. The Selection of B-Line

The breeding goal of the B-line was to obtain a fertile line, which was similar to BTx623 and carried the SB6 waxy mutation (wx^c allele). This line will provide the pollen for use in the A-line to maintain its male sterile characteristics along with the wx^c allele (Figure 6). In the first cropping season in 2015, SB6 was used as the female parent and BTx623 as the male parent for artificial hybridization through emasculation and pollination. This resulted in the establishment of F₁ hybrid combinations, yielding 24 hybrid F₁ seeds. In the second cropping season in 2015, 24 F₁ plants were grown as female parents, with BTx623 as the male parents, for artificial backcrossing. This led to the production of 80 BC₁F₁ seeds.

In the following year, in the first cropping season in 2016, 215 BC₁F₁ plants were obtained, and during their growth, those with favorable growth characteristics and similarity to BTx623 were selected for wx^c allele selection. Out of these, 35 plants exhibited heterozygous combinations of the wx^c allele (Wx^c/wx^c) and were subjected to artificial backcrossing with BTx623. This resulted in the generation of 200 BC₂F₁ backcross seeds. BC₂F₁ plants were grown, and in the second cropping season in 2016, 72 plants showing heterozygous combinations of the wx^c allele (Wx^c/wx^c) and having favorable growth characteristics similar to BTx623 were selected for further artificial backcrossing.

The result was the production of BC₃F₁ backcross seeds, which were planted in the first cropping season in 2017. The BC₃F₁ plants underwent wx^c allele selection, leading to the selection of 96 plants exhibiting heterozygous combinations of the wx^c allele (Wx^c/wx^c). These 96 waxy plants were transplanted into the field and subjected to agricultural trait selection and bagged self-pollination during the flowering period. In the second cropping season in 2017, approximately 210 BC₃F₂ plants were subjected to wx^c allele selection. Among these, 45 plants exhibited homozygosity of the wx^c allele (wx^c/wx^c). These 45 waxy plants were transplanted into the field and selected during flowering, resulting in 20 plants chosen for bagged self-pollination. Starting in 2018, the 20 BC₃F₃ plants were mixed to form a fertile maintenance parent population. Inferior single plants were eliminated, and three to five plants with characteristics similar to BTx623 were selected. In 2021, superior fertile maintenance lines were designated as CNB1 (Figure 7).

3.4. The Selection of R-Line

The breeding goal of the R-line was to obtain a restorer similar to BTx623, possessing the wx^c allele, and when crossed with the A-line, the F₁ hybrid could restore fertility (Figure 8). In the first cropping season in 2015, F₅ waxy lines from TCR170617/SB6, which could restore fertility, were used as the female parent, with BTx623 as the male parent, for artificial hybridization through emasculation and pollination. This resulted in 38 hybrid F₁ seeds. In the second cropping season in 2015, the F₁ plants were grown as female parents, with BTx623 as the male parents, for artificial backcrossing. This led to the production of 90 BC₁F₁ seeds.

In the following year, in the first cropping season in 2016, 90 BC₁F₁ plants were obtained, and during their growth, those with favorable growth characteristics and similarity to BTx623 were selected for wx^c allele selection. Out of these, 30 plants exhibited heterozygous combinations of the wx^c allele (Wx^c/wx^c) and were subjected to artificial backcrossing with BTx623. This resulted in the generation of 200 BC₂F₁ backcross seeds. BC₂F₁ plants were grown, and in the second cropping season in 2016, 96 plants showing heterozygous combinations of the wx^c allele (Wx^c/wx^c) and having favorable growth characteristics similar to BTx623 were selected for further artificial backcrossing.

The result was the production of BC₃F₁ backcross seeds, which were planted in the first cropping season in 2017. The BC₃F₁ plants underwent wx^c allele selection, leading to the selection of 62 plants exhibiting heterozygous combinations of the wx^c allele (Wx^c/wx^c). These 62 waxy plants were transplanted into the field and subjected to bagged self-pollination during the flowering period. In the second cropping season in 2017, approximately 100 BC₃F₂ plants were subjected to wx^c allele selection. Among these, 26 plants exhibited homozygosity of the wx^c allele (wx^c/wx^c). These 26 waxy plants were transplanted into the field and selected during flowering, resulting in 20 plants being chosen for bagged self-pollination. Starting in 2018, the 20 BC₃F₃ plants were mixed to form a fertile restoration parent population. Inferior single plants were eliminated, and three to five plants with characteristics similar to BTx623 were selected. In 2021, superior fertility restorer lines were marked as CNR1 (Figure 9).

3.5. Agronomy Traits Analysis

3.5.1. Preliminary Trials

The agronomic characteristics of three waxy sorghum lines—A-, B-, and R-lines—were investigated in the first cropping season in 2022 (

Table 1. The agronomic traits and preliminary yield trial among CNA1, CNB1, and CNR1.

Line	Heading Days	Days to Harvesting	Height (cm)	Panicle Length (cm)	Panicle Weight (g)	Number of Seeds per Panicle	Seed Weight per Panicle (g)	Yield (kg ha−1)
CNA1	60 a	111 a	148.5 ± 1.0 a	31.1 ± 1.5 a	69.61 ± 3.0 b	1872 ± 38.6 c	51.70 ± 0.2 b	2913 ± 9.3 b
CNB1	56 b	106 b	151.9 ± 2.7 b	31.2 ± 0.4 a	96.12 ± 5.0 a	3335 ± 11.0 a	78.70 ± 2.8 a	4434 ± 154.9 a
CNR1	53 c	97 c	132.5 ± 0.4 c	30.9 ± 0.2 a	91.95 ± 2.6 a	2951 ± 95.2 b	74.68 ± 2.1 a	4310 ± 118.3 a

Each value represents the means of three replicates ± standard deviation. Different Latin superscript letters denote groups, which differ significantly (p < 0.05).

). The heading days results showed that CNA1 had the latest heading date, followed by CNB1, and CNR1 had the earliest heading date. Regarding plant height, CNB1 had the tallest plant height, followed by CNA1, and CNR1 had the shortest plant height.

Regarding panicle length, the three lines had similar panicle lengths, with no significant differences observed. CNA1 had the smallest panicle weight. CNB1 and CNR1 had larger panicle weights, and there was no significant difference between the two. In terms of the number of seeds per panicle, CNB1 had the most seeds, followed by CNR1, and CNA1 had the fewest. The seed weight per panicle and panicle weight showed similar results, with CNA1 being the smallest, CNB1 and CNR1 being larger, and no significant differences being observed between the two. The yield was also the lowest for CNA1, while CNB1 and CNR1 yielded more, with no significant differences being observed between them (Table 1).

3.5.2. Advanced Trials

The agronomic characteristics of the three waxy sorghum lines (A-, B-, R-lines) and F₁ hybrid (CNH1) were investigated in the second cropping season in 2022 (

Table 2. The agronomic traits and advanced yield trial among CNA1, CNB1, CNR1, and CNH1 (F₁).

Line	Heading Days	Days to Harvesting	Height (cm)	Panicle Length (cm)	Panicle Weight (g)	Number of Seeds per Panicle	Seed Weight per Panicle (g)	Thousand-Seed Weight (g)
CNA1	59.75 ± 2.4 a	109.00 ± 2.0 a	110.36 ± 1.4 c	31.95 ± 1.4 a	28.54 ± 6.7 c	398.4 ± 341 c	8.58 ± 7.6 c	20.82 ± 1.5 b
CNB1	60.50 ± 2.3 a	107.00 ± 2.4 a	115.68 ± 4.6 a	29.85 ± 1.4 c	109.22 ± 33.4 a	3327.1 ± 1013 a	83.86 ± 26 a	25.15 ± 1.6 a
CNR1	56.75 ± 0.4 a	104.75 ± 2.9 a	102.45 ± 3.7 d	30.74 ± 1.4 bc	106.96 ± 27.4 a	3137.3 ± 708 a	85.46 ± 22.9 a	27.08 ± 1.6 a
CNH1	59.00 ± 2.5 a	106.00 ± 1.0 a	112.71 ± 2.3 b	30.90 ± 0.86 b	88.40 ± 22.3 b	2687.8 ± 545 b	71.46 ± 19.6 b	26.41 ± 1.9 a

Each value represents the means of three replicates ± standard deviation. Different Latin superscript letters denote groups, which differ significantly (p < 0.05).

). Due to the differences in sunlight, temperature, and rainfall between Taiwan’s first and second crop seasons, the days to heading for CAN were not significantly different. However, for CNB1 and CNR1 in the second crop season, the days to heading were longer compared to the first crop season. CNH1 had a heading time in the second crop season, which was similar to CNA1 and CNB1. In terms of days to harvesting, CNB1 showed no significant difference between the first and the second crop seasons, while CNA1 had a slightly shorter duration in the second crop season compared to the first crop season. CNR1, on the other hand, had a longer harvest duration in the second crop season compared to the first crop season. However, all four varieties (CNA1, CNB1, CNR1, CNH1) had very similar days to harvesting.

In terms of plant height, there was a significant difference among the four, with CNB1 being the tallest and CNR1 being the shortest. There was a noticeable difference in plant height between the first and the second crop seasons, with the first crop season having taller plants and the second crop season having shorter ones, but CNB1 remained the tallest while CNR1 remained the shortest in both seasons. As for panicle length, there was little difference among the four, with CNA1 having the longest panicle and CNB1 having the shortest.

Regarding panicle weight, the number of seeds per panicle, and seed weight per panicle, both CNB1 and CNR1 had the highest values, and there was no significant difference between them. CNA1, on the other hand, had the lowest values, with CNH1 falling in between. In terms of thousand-seed weight, CNB1, CNR1, and CNH1 were heavier, and there was no significant difference among the three, while CNA1 was lighter. In the second crop season experiment of CNA1, there was a significant performance gap in terms of panicle weight, grains per panicle, grain weight per panicle, and thousand-seed weight due to poor pollination.

4. Discussion

This study successfully bred sorghum materials using the three-line breeding system, incorporating genes for stickiness (waxy) into the sorghum lines. The recurrent parents in the breeding cycle were all based on the sorghum reference variety BTx623, which has a fully sequenced genome, making it a valuable genetic resource for future genetic and genomic research [26]. Additionally, this study introduced the stickiness gene wx^c from Taiwan’s native sticky sorghum variety SB6. The research conducted initial yield comparison trials among the three components of the three-line system (A-line, B-line, R-line), evaluating the yield and agronomic traits. Preliminary results were obtained, and in the same year, advanced comparison trials were conducted, including the assessment of F₁ hybrid offspring, to further compare the differences among these components.

The introduction of the waxy gene into the sorghum lines and the utilization of the three-line breeding system hold significant promise for enhancing sorghum crops with the desired traits. The combination of advanced genetic materials and rigorous testing methodologies opens the door to improved sorghum varieties, which can address the evolving needs of agriculture and contribute to food security [27].

The development of this three-line breeding system for sorghum, which incorporates stickiness genes, spanned eight years, from 2015 to 2022. Through multiple rounds of backcrossing and self-pollination, this breeding effort resulted in valuable experimental materials for genetic and breeding research worldwide [28]. In terms of industry, it holds immense promise, particularly benefiting the liquor production sector in Taiwan, where most of the sorghum used for brewing is imported. With the availability of this new material, there is an opportunity to enhance the value of the brewing industry, which is excellent news for both farmers and businesses alike. This innovative approach not only represents a significant achievement in genetic and agricultural research, but it also has the potential to transform the sorghum-based liquor production landscape [29,30].

5. Conclusions

In this study, the breeding outcomes demonstrated that the utilization of backcross breeding, molecular-assisted selection, and traditional field selection methods successfully created three-line materials (referred to as CNA1, CNB1, and CNR1 for A-, B-, and R-lines, respectively) and an F₁ hybrid (CNH1) with desirable agronomic traits and yields. These achievements are the result of years of dedicated research and effort, establishing a robust basis for future advancements in both the scientific and industrial realms of sorghum breeding and utilization.