The seeds of the soybean (Glycine max (L.) Merr.) contain 40% protein, 20% oil, and 15% soluble carbohyd-rates, making soybean one of the world's most economi-cally important crops. The global soybean production in 2021 was 367.8 million metric tons, of which 164.5 million metric tons were exported from soybean-producing coun-tries to soybean-importing countries (Soystat 2022, http://soystats.com/). The soybean yield is affected by environmental conditions, genotype, and crop management, and is one of the most important breeding goals (Ball et al. 2000). The number of pods per plant, number of seeds per pod, seed weight, and number of plants per unit are important components for improving soybean yield. Since the 1930s, researchers have studied the number of seeds per pod, with a focus on four-seeded pods, to increase soybean yield (Takahashi 1934; Domingo 1945; Li et al. 2018).

Although the number of seeds per pod is a complex quantitative trait in soybeans, it is highly heritable in successive generations (Wang et al. 2007; Pei et al. 2010). The number of seeds per pod is depending on the number of ovules per placenta. Generally, two- and three-seeded pods are common in soybean; only a few cultivars have five-seeded pods (Zhu and Sun 2006). The increased ratio of four-seeded pods per plant is closely associated with soybean yield. The number of four-seeded pods is positively correlated with the number of pods per plant, which is also positively associated with soybean yield (Peng et al. 1994; Li et al. 2018).

To date, several studies with linkage analyses have been conducted to identify quantitative trait locus (QTL) for four-seeded pods in soybean, using different background of genetic populations. Six stable QTLs for the number of four-seeded pods were detected in 11 different environments (Asakura et al. 2012; Yang et al. 2013). Zhou et al. (2009) detected three QTL for the number of four-seeded pods with 165 recombinant inbred lines (RILs). Gao et al. (2012) reported that eight QTL on chromosomes 1, 5, and 12 were responsible for the number of four-seeded pods in an RIL population. Recently, Li et al. (2021), utilizing QTL mapping and RNA-seq with chromosome segment substi-tution lines (CSSLs), identified Glyma.07G200900 and Glyma.07G201200 as being involved in cell division and, consequently, the number of four-seed pods in soybean.

Soybean plants with lanceolate leaves were found to have a higher number of four-seeded pods per plant than did plants with oval leaves, suggesting that alleles controlling narrow leaflets had pleiotropic effects on the number of four-seeded pods in soybean or were tightly linked to the number of four-seeded pods (Takahashi 1934; Domingo 1945; Johnson and Bernard 1962; Weiss 1970). Thus, a narrow leaflet can be used as a morphological marker for the four-seeded pods of soybean. Zhu and Sun (2006) showed a 3:1 segregation ratio of the F₂ population for four-seeded pods per plant in soybean, indicating that this is controlled by a single recessive gene. Jeong et al. (2012) identified the Ln gene (Glyma.20G116200) on chromosome 20 for leaflet shape and number of seeds per pod, that it encoded GmJAGGED1, which is a homolog gene to regulates lateral organ development in Arabidopsis (Dinneny et al. 2004; Ohno et al. 2004). The results of a previous study revealed that the lnln genotype had lanceolate leaves and a higher number of four-seeded pods than plants with oval leaves (LnLn), with the result that the Ln gene had pleiotropic effects for the number of four-seeded pods and leaf shape (Dinkins et al. 2002).

The number of seeds per pod at each node was measured to understand the effect of position on the number of seeds per pod in soybean (Khare 2011). The authors reported that the highest number of four-seeded pods was on the second node from the top, whereas the lowest was on the basal node. In addition, four- and three-seeded pods showed a 1:1 ratio (on an overall plant basis) in a progeny line, JS 90-41, derived from an interspecific crossing population. It was found that this soybean genotype showed a 3:1 ratio of four- to three-seeded pods at one-third of the distal part of soybean and a 1:3 ratio of four- to three-seeded pods at one-third of the basal part of soybean (Khare 2011).

Mutation breeding with chemical mutagens has been successfully increased novel genetic diversity in soybeans (Carrol et al. 1985; Hoffman et al. 1999; Kim et al. 2015; Jeong et al. 2018; Thapa et al. 2018; Jo et al. 2021). The benefit of mutation breeding is that mutations in the genome are heritable in subsequent generations of crops. In our previous study, we have used ethyl methanesulfonate (EMS) as chemical mutagen to select soybean mutant line (PE2248) from a mutation population (Chae et al. 2013). The objective of the present study was to investigate the agronomic traits and number of seeds per pod of a useful genetic material, FS1159, derived from the cross of Jinpung and a progeny line from the F₃ generation of Uram × PE2248. The following traits were measured in the evaluated accessions: plant height, number of nodes, number of pods, 100-seed weight, yield, number of seeds per pod, and number of four-seeded pods per plant.

Plant materials and growth environment

To evaluate the agronomic traits and number of seeds per pod, a progeny line, FS1159, with an increased ratio of four-seeded pods, and four check soybeans, namely Jinpung (Lee et al. 2015), Uram (Ko et al. 2016), PE2248, and Pungsannamul (Lee et al. 2015), were used in this study. Jinpung, Uram, and Pungsannamul are Korean soybean cultivars developed by Rural Development of Administration, Republic of Korea and PE2248 was a mutant line derived from 0.3% EMS treatment of Pungsan-namul (Chae et al. 2013). The F_4:6 of FS1159 was derived from crossing Jinpung and a progeny line from the F₃ generation of Uram × PE2248.

The soybean genotypes were planted in an experimental field affiliated with Kyungpook National University (Gunwi, 36°07′N, 128°38′E). The planting dates were May 27, June 10, and June 24, 2021. The experimental design was a randomized complete block (RCB) with two replications. Each plot consisted of three rows, 2 m long and spaced 70 cm apart. Seeds were planted manually in hills within rows spaced 15 cm apart from each other. Plants were thinned to a single stand per hill at the V2-3 stage. Five randomly selected plants from the center row of each three-row plot at maturity (R8) (Fehr et al. 1971) were used to measure the plant height, number of branches per plant, number of nodes per plant, 100-seed weight, and number of pods per plant. Plant height was measured from the base of the stem at ground level to the tip of the top pod. All pods from five plants from each plot were harvested for further assessment at full maturity (Fig. 1). The number of one-, two-. three-, and four-seeded pods per plant were averaged for five plants from each plot.

Figure 1. Seed number per pod with four soybean genotypes. (A) PE2248, (B) Pungsannamul, (C) Jinpung, and (D) FS1159.

Statistical analysis

Statistical analyses in this study were conducted using SAS v9.4 (SAS Institute 2013). Analysis of variance (ANOVA) was conducted to evaluate differences for genotype, planting date and genotype by planting date interaction using the PROC GLM in SAS. Mean differences among the genotypes with measurements of agronomic traits were analyzed by applying Fisher’s least significant difference (LSD) test at P ≤ 0.05, using PROC GLM.

An ANOVA was conducted to determine the influence of environmental and genetic effects on the agronomic traits (Table 1). Plant height and 100-seed weight were significantly associated with the genotypic effects (P < 0.001). The planting date significantly affected plant height, number of nodes, number of branches, and total seed weight. Plant height was the only trait that showed significant genotype × planting date interaction (P < 0.01). A larger mean square value indicated a greater influence on each agronomic trait. The primary factors for the number of nodes, number of branches, and total seed weight were environmental factors, with 52.5%, 79.3%, and 44.8% of the total mean squares, respectively. However, the values of the genotypic effect on plant height and 100-seed weight were 46.9% and 93.6% of the total mean squares, respectively.

Table 1 . Mean square from analysis of variance (ANOVA) of each measured agronomic trait and seed weight of four soybean genotypes evaluated from three planting dates.

Source of variation	Degree of freedom	Plant height	Number of nodes per plant	Number of branches per plant	Number of pods per plant	100-seed weight	Number of total seeds per plant	Total seed weightz)
Genotype (G)	4	229.5***	8.0	1.4	4549.6	354.7***	23843.9	2390.5
Planting date (PD)	2	158.8***	16.7*	13.4***	6233.6	4.9	21926.3	5736.1*
Replication in PD	3	20.45	0.2	0.1	2699.9	3.8	12336.3	945.3
G × PD	8	69.6**	3.9	1.1	2608.4	9.4	13024.9	2479.1
Error	12	11.4	3.0	0.9	2064.6	6.1	9402.1	1256.7

***Significant at the 0.001 probability level. **Significant at the 0.01 probability level. *Significant at the 0.05 probability level.

^z)Averaged total seed weight of five single plants from each plot.

The following traits of the five soybean genotypes were evaluated: plant height, number of nodes, number of branches, number of pods, 100-seed weight, number of total seeds, and total seed weight (Table 2). Plant height was significantly greater in FS1159 (58.1 cm) and Uram (53.5 cm) than in Jinpung (46.1 cm), PE2248 (43.7 cm), and Pungsannamul (45.1 cm). The number of nodes in FS1159 was15.7. The mean number of branches in FS1159 (6.5) was not significantly different from that of the other genotypes. The 100-seed weight of FS1159 was 20.3 g, which was significantly higher than the corresponding values in Pungsannamul (13.7 g) and PE2248 (13.8 g).

Table 2 . Agronomic traits and seed weight of four soybean genotypes evaluated from three planting dates.

Genotype	Plant height (cm)	Number of nodes per plant	Number of branches per plant	Number of pods per plant	100-seed weight (g)	Number of total seeds per plant	Total seed weight (g)z)
FS1159	58.1a,y)	15.7a	5.6a	109.0ab	20.3b	261.7abc	149.5ab
Jinpung	46.1c	13.0b	6.5a	101.5b	29.3a	195.8bc	155.3a
PE2248	43.7c	14.8ab	5.9a	159.0a	13.8c	324.0a	110.2b
Pungsannamul	45.1c	15.9a	6.8a	149.5ab	13.7c	303.3ab	121.0ab
Uram	53.5b	13.3b	5.8a	93.7b	30.5a	166.8c	137.5ab

^y)The same letter within each agronomic trait indicates no significant difference base on least square difference (P < 0.05).

^z)Averaged total seed weight of five single plants from each plot.

The relative numbers of one-, two-, three-, and four-seeded pods for the five soybean genotypes are shown in Fig. 2. The portion of four-seeded pods relative to the total number of pods on different planting dates ranged from 0 to 11.3% in the different soybean genotypes. The soybean genotype, FS1159, contained a significantly higher ratio of four-seeded pods (11.3%) than did other genotypes (range of 0.0-0.7%). Overall, the largest portion of pods in FS1159 was three-seeded pods (39.6%). Generally, two-seeded and three-seeded pods are common in soybeans. Only a few cultivars have five seeded pods (Zhu and Sun 2006). In this study, the highest number of seeds per pod was two-seeded pods in Jinpung (50.7%), PE2248 (49.7%), Punsannamul (43.1%), and Uram (52.4%). Khare (2011) reported that four- and three-seeded pods occurred at a 1:1 ratio (overall plant basis). However, FS1159 showed an approximately 1:4 ratio of four-to three-seeded pods, whereas it showed an approximately 1:1 ratio of three- to two-seeded pods in this study.

Figure 2. Comparison of relative values for the ratio of number of seeds per pod with all soybean geno-types. The values in the bar represent mean values of each genotype measured in different planting dates. The same letter within each ratio of number of seeds per pod indicates no significant difference at P < 0.05.

Researchers have reported that narrow leaflet soybean with an increased ratio of four-seeded pods consistently showed smaller seeds than did broad-leaflet plants (Weiss 1970; Mandl and Buss 1981). FS1159 can be a unique genetic or breeding material that has both a medium seed size and a higher ratio of four-seeded pods per plant. We assumed that the trait of the four-seeded pod in FS1159 may be associated with additive genetic effects or more than two candidate genes, given that FS1159 was from a crossing combination with three different soybean genotypes: PE2248, Jinpung, and Uram. Therefore, because the number of four-seeded pods has been a trait of interest for improving soybean yield, it is of great value to identify linked markers or gene-based markers for the phenotype of four-seeded pods; QTLs mapping studies, and markers can be used for marker-assisted selection in soybean breeding programs.

In summary, our results show that FS1159 can be used as a new genetic and breeding material to further understand the four-seeded pod trait in soybean as well as to improve soybean yield.

This work was carried out with the support of the Cooperative Research Program for Agriculture Science and Technology Development (Project No. PJ01416803), Rural Development Administration, Jeonju, Republic of Korea.