This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Tofu, a widely consumed soy-based product, has seen increasing consumer demand, driven by health and dietary trends. Enhancing the processing suitability of soybeans requires analysis of the genetics associated with tofu quality traits. In this study, five key traits, namely water absorption (WA%), Brix of soymilk (°BS), soymilk yield (YM%), tofu residue (TR%), and tofu yield (YT%), were evaluated in two recombinant inbred line (RIL) populations: 81 lines of “Daepung”דSaedanbaek” and 122 lines of “Daepung”דGalchae”. Most traits exhibited normal distributions and transgressive segregation. Trait correlations revealed strong associations between WA and TR, and between YM and YT, with population-specific variation. Quantitative trait locus (QTL) mapping identified eight loci in each population linked to WA, °BS, YM, TR, and YT across multiple chromosomes. Notably, co-localized QTL were detected for WA% and TR% (chromosomes 1 and 6) and for WA%, YM%, °Brix, and TR% (chromosome 8), indicating pleiotropic regulation. Candidate genes included Glyma.06g173500, associated with β-glucan biosynthesis, and Glyma.03g132700, encoding β-glucanase, both implicated in seed coat permeability and tofu yield. These results provide novel genetic insights into seed coat properties and processing quality, and the identified QTLs and candidate genes represent valuable targets for marker-assisted selection in soybean breeding programs optimized for tofu production in Korea.
Soybean (Glycine max L.) is a globally important legume crop, valued for its high protein (40%) and oil (approximately 18%) contents, which make it essential for food, feed, and industrial applications (Modgil et al. 2021). In addition to proteins, soybean contains lipids, vitamins, minerals, free sugars, and bioactive constituents such as isoflavones, flavonoids, saponins, and peptides, many of which exhibit therapeutic properties (Kim et al. 2006, Park et al. 2016). Soybean is consumed in unfermented forms such as roasted and fried beans, soy powder, soybean oil, and soy butter, as well as in fermented products including soy sauce, soybean cheese, pickles, and soy yogurt (Sanjukta & Rai 2016).
In Korea, soybean consumed in diverse processed forms including soymilk, fermented pastes (doenjang, cheonggukjang, gochujang, meju), soy yogurt and tofu (Oh et al. 1981, Qin et al. 2022). Among these, tofu is rich in high-quality protein, calcium, iron, and essential amino acids, making it a valuable supplement to cereal-based diets (Jung & Kim 2016). Beyond its nutritional value, tofu contains bioactive compounds such as saponins, flavonoids, and lecithin, which are associated with antioxidant, anticancer, and cognitive health benefits (Jung & Kim 2016).
Tofu quality and yield are strongly influenced by soybean seed composition, especially the ratio of 7S (β-conglycinin) to 11S (glycinin) globulins, as well as processing factors such as soaking, heating, coagulant concentration, and pressing methods (Guan et al. 2021, Hou et al. 2025, Kinsella 1979). In Korea, although high-protein cultivars such as ‘Saedanbaek’ have been developed to enhance tofu quality, their relatively low seed yield has limited their adoption. Conversely, cultivars like ‘Galchae’ show high soybean residue content and reduced tofu yield despite favorable agronomic traits, highlighting the trade-offs breeders must balance (Jung et al. 2022).
Recent advances in soybean genomics, including whole-genome sequencing, high-density SNP arrays, and high-throughput genotyping platforms, have facilitated QTL mapping and GWAS for key agronomic traits such as seed protein content, seed oil content, seed weight, flowering time, and shattering resistance (Jo et al. 2022, Liu et al. 2023, Patel et al. 2025, Patil et al. 2025, Petereit et al. 2022).
In addition, recent studies have begun to investigate the genetic basis of tofu quality traits in soybean (Döttinger et al. 2025, Kurasch et al. 2018). Döttinger et al. (2025) identified QTLs associated with important tofu attributes, including tofu yield and hardness, highlighting the potential for marker-assisted selection to improve processing quality. However, studies specifically targeting tofu-related traits remain limited. Moreover, conventional tofu evaluation methods require large seed quantities, which limits their applicability in early-generation breeding populations. Therefore, simplified evaluation methods are necessary to efficiently assess tofu quality and accelerate the identification of QTLs and candidate genes associated with these traits (Jung et al. 2022, Kim et al. 2021).
In this study, we focused on Korean soybean and aimed to evaluate key tofu quality traits, including water absorption (WA%), soymilk Brix (°BS), soymilk yield (YM%), tofu residue (TR%), and tofu yield (YT%), and to identify associated QTLs and candidate genes. The ultimate goal is to provide genetic insights into soybean processing suitability and to support marker-assisted selection for the development of high-quality tofu cultivars in Korea.
Materials and Methods
Plant Materials and Population Development
In this study, the simplified centrifugation-based method for evaluating tofu quality traits, previously developed and validated by our group (Jung et al. 2022), was applied to genetic populations for QTL analysis. Among the 20 soybean cultivars initially tested in the previous study, ‘Saedanbaek’ (exhibiting the highest YT) and ‘Galchae’ (exhibiting the lowest YT) were selected as contrasting parents for population development. The high-yielding cultivar ‘Daepung’ was used as the common female parent (Park et al. 2005), while ‘Saedanbaek’ (Kim et al. 2014), a processing-type cultivar, and ‘Galchae’ (Oh et al. 2009), a small-seeded sprout-type cultivar, served as the male parents. Two recombinant inbred line (RIL) populations, advanced by the single seed descent (SSD) method, comprised F6:9 lines from the Daepung×Saedanbaek and Daepung×Galchae crosses.
Traits Evaluation and Statistical Analyses
Two RIL populations were cultivated in Cheonan, Chungcheongnam-do, Republic of Korea (36°43′N, 127°16′E). Standard agronomic practices and appropriate plant protection measures were applied to ensure optimal crop growth and development across all stages. The quality traits evaluated included WA%, °BS, YM%, TR%, and YT%. Tofu production and related quality parameters, including water absorption (WA%), soymilk °Brix (°BS), soymilk yield (YM%), tofu residue (TR%), and tofu yield (YT%), were evaluated following ‘the fast and simple processing method to evaluate tofu’ described by Jung et al. (2022) For water absorption (WA%), 15 g of soybeans were soaked in water for 12 hours, and the weight before and after soaking was measured to determine the percentage increase. Soymilk °Brix (°BS) was measured using a digital refractometer (PAL-1, ATAGO, U.S.A) to determine the soluble solids content of the prepared soymilk. Soymilk yield (YM%) was calculated as the percentage ratio of the weight of the filtered soymilk to the total weight of raw soybeans and water used during processing. For tofu residue (TR%), the residue remaining after soymilk extraction was dried at 70℃ for 3 days in a dry oven (NB-901M, N-BIOTEK, Korea), and the percentage of dried residue weight per unit weight of soybeans was calculated. Finally, tofu yield (YT%) was determined by measuring the dried weight of the molded tofu and calculating its percentage relative to the weight of the raw soybeans used.
Statistical analyses were performed using RStudio (version 4.0.2; RStudio, Inc.). The data distribution for normality was evaluated using the Shapiro-Wilk test (Shapiro & Wilk 1965). In addition, histograms with overlaid normal distribution curves and kernel density estimates were generated using the R package ggplot2. Trait relationships were examined through correlation analysis, and statistical significance was determined accordingly.
DNA Extraction and Genotyping
Genomic DNA was isolated from young soybean leaves following a modified cetyltrimethylammonium bromide protocol (Doyle & Doyle 1987). DNA quantity and quality were assessed using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). Genotyping of the RILs was performed using the Axiom® 180K SoyaSNP array (Affymetrix, CA, USA), which contains 180,961 SNP markers developed for soybean (Lee et al. 2015).
Linkage Map Construction and QTL Mapping
The linkage map was constructed using IciMapping version 4.2 (Meng et al. 2015), based on SNP markers that were polymorphic in the ‘Daepung’בSaedanbaek’ and ‘Daepung’×Galchae RIL populations. QTL analysis was performed using the ICIM (Inclusive Composite Interval Mapping) method, with a LOD threshold of ≥2.5 for significant QTLs after. Marker ordering was conducted using the nnTwoOpt algorithm, and map rippling was performed with the SARF criterion. Significant QTLs were identified through a permutation test with 1,000 iterations at p≤0.05.
Candidate Genes Identification
Candidate genes within the QTL regions were identified based on their physical positions, using the ‘Soybean Wm82a2v1’ reference genome available from the Soybean genomic database, SoyBase (http://www.soybase.org). Functional annotation of the putative genes was performed using in silico analysis (https://www.uniprot.org/).
Results and Discussion
Phenotypic Evaluation of Tofu Quality Traits
We evaluated genetic variation for tofu quality traits in two RIL populations: 81 F6:9 RILs from Daepung×Saedanbaek and 122 F6:9 RILs from Daepung×Galchae. In the Daepung×Saedanbaek population (Supplementary Table 1), WA% ranged from 206.3% to 242.0% (mean 227.4%), °BS from 6.1% to 7.7% (mean 6.9%), YM% from 80.3% to 89.3% (mean 86.6%), TR% from 14.4% to 23.9% (mean 18.5%), and YT% from 47.0% to 57.3% (mean 50.7%). Parental means were within these ranges. In the Daepung×Saedanbaek population, WA% and YM% exhibited deviations from normality, with WA% showing a slightly skewed distribution and YM% displaying stronger departure from normality. In contrast, °BS, TR%, and YT% followed normal distributions (Fig. 1), confirming the suitability of this population for QTL analysis and breeding.
In the Daepung×Galchae population (Supplementary Table 2), WA% ranged from 153.8% to 240.3% (mean 228.3%), °BS from 5.7% to 7.6% (mean 6.9%), YM% from 80.9% to 89.5% (mean 86.1%), TR% from 15.3% to 26.4% (mean 20.2%), and YT% from 46.6% to 55.3% (mean 51.3%). Trait distributions for WA%, TR%, and YT% did not deviate from normality, whereas °BS and YM% showed significant departures from normality (Fig. 2). This indicates that the Daepung×Galchae population also represents valuable breeding material. Notably, Galchae contributed alleles associated with high residue content and low tofu yield, consistent with its poor performance in the simplified processing evaluations reported by (Jung et al. 2022). However, several RILs exceeded both parental values, demonstrating clear evidence of transgressive segregation and highlighting the potential for genetic improvement of tofu quality. These results are in agreement with findings by Döttinger et al. (2025), who also reported transgressive segregation for all tofu quality traits and emphasized that such variation, even when parental means fall below commercial thresholds, is critical for successful QTL mapping and breeding progress in soybean.
The correlation analysis of tofu quality traits in the Daepung×Saedanbaek RIL population (Table 1) revealed that WA% was most strongly correlated with TR% (r=0.557, p<0.001), followed by YT% (r=0.228, p<0.01) and °BS (r=0.222, p<0.001). No significant correlation was observed between WA% and YM% (r=0.041). °BS showed a significant positive correlation with YT% (r=0.354, p<0.001), whereas its correlations with YM% (r=-0.201) and TR% (r=-0.004) were not significant. Similarly, YM was not significantly correlated with either TR% (r=-0.096) or YT% (r=0.037). In contrast, TR% was significantly and positively correlated with YT% (r=0.159, p<0.05). In the Daepung×Galchae RIL population (Table 1), WA% showed a positive correlation with YM% (r=0.248, p<0.001) but was not significantly associated with °BS (r=0.006), TR (r=0.033), or YT% (r=0.066). °BS was negatively correlated with TR% (r=-0.182, p<0.01) and positively correlated with YT% (r=0.329, p<0.001), while its correlation with YM% was not significant (r=0.083). YM% exhibited a negative correlation with TR% (r=-0.186, p<0.01) and a positive correlation with YT% (r=0.416, p<0.001). TR% was negatively correlated with YT% (r=-0.192, p<0.001).
Phenotypic correlation analysis of tofu quality traits in the two RIL populations revealed distinct patterns. In the Daepung×Saedanbaek population, the strongest correlation was detected between WA% and TR% (r=0.557), whereas in the Daepung×Galchae population, the strongest correlation was observed between YM% and YT% (r=0.416). A consistent feature across both populations was the positive correlation between °BS and YT%. Despite being evaluated under identical experimental conditions, the two populations exhibited different correlation structures, suggesting that divergent outcomes may also be anticipated in subsequent QTL detection.
Our findings align with those of Döttinger et al. (2025), who also reported substantial trait interdependencies among tofu quality parameters. Notably, they observed a strong positive correlation between tofu yield (TY) and tofu weight (TW) (r=0.87) and a strong negative correlation between TY and tofu hardness (TH) (r=-0.65). These results reinforce the challenge of simultaneously improving tofu yield and texture, as higher yield is typically linked to increased water content, whereas firmer tofu tends to retain less water. Both studies also provide evidence for transgressive segregation and demonstrate that parental lines contribute contrasting alleles that can be recombined to generate progeny with superior performance. Together, these findings highlight the complexity of the genetic architecture underlying tofu quality traits and underscore the need to balance correlated attributes when selecting for consumer-preferred tofu varieties.
Construction of Genetic Maps
Genotyping of the two RIL populations was performed using the 180K Axiom® SoyaSNP array, which includes 180,961 markers, and genetic maps were constructed with ICIM 4.2. The raw genotyping data for the two RIL populations (K-SoyNAM) are publicly available in the figshare Dataset Catalog (Kim 2024). In the Daepung×Saedanbaek population (Supplementary Table 3), a total of 7,922 SNP markers were mapped. The number of markers per chromosome ranged from 49 to 638, with an average of 396 markers. The resulting genetic map spanned 2,035.73 cM, with an average length of 101.79 cM per chromosome. Among the 20 soybean chromosomes, chromosome 12 was the shortest (48.52 cM), while chromosome 20 was the longest (132.12 cM). The average marker interval was 0.26 cM, indicating a high-density genetic map. In the Daepung×Galchae population (Supplementary Table 4), 16,639 SNP markers were mapped, with 476 to 1,382 markers per chromosome and an average of 831.95 markers. The total genetic map length was 2,769.92 cM, with an average of 138.50 cM per chromosome. Chromosome 16 was the shortest (91.54 cM), whereas chromosome 18 was again the longest (212.57 cM). The average inter-marker distance was 0.17 cM, reflecting a higher marker density compared with the Daepung×Saedanbaek population.
QTL Mapping
QTL mapping of tofu quality traits in the Daepung×Saedanbaek RIL population identified eight QTLs. No QTL was detected for YM%. Three QTLs for WA% were located on chromosomes 1, 6, and 20, with LOD scores of 4.09, 5.83, and 6.11, explaining 13.0%, 16.1%, and 17.4% of phenotypic variance, respectively. The additive effects were 0.508, -0.55, and -0.56, indicating that Daepung allele increased WA on chromosome 1, whereas Saedanbaek allele contributed positively on chromosomes 6 and 20. One QTL for °BS was detected on chromosome 17 (LOD 2.94; PVE 19.8%; Add -0.13). Three QTLs for TR% were identified on chromosomes 1 (two QTLs) and 6, with LOD scores of 2.99 and 8.82 on chromosome 1 (PVE 6.3% and 22.4%; Add -0.69 and 1.30) and 7.13 (PVE 16.4%; Add -1.08) on chromosome 6. One QTL for YT% was detected on chromosome 3 (LOD 2.78; PVE 12.7%; Add -0.12). Notably, WA% and TR% QTLs co-localized at the same marker positions on chromosomes 1 and 6, suggesting potential pleiotropic effects (Table 2, Fig. 3). QTL mapping of tofu quality traits in the Daepung×Galchae RIL population identified eight QTLs. No QTL was detected for YT%. Two QTLs for WA% were located on chromosomes 8 and 15 with LOD scores of 2.82 and 3.79, explaining 10.5% and 15.0% of phenotypic variance, respectively, with additive effects of 1.55 and 1.85, indicating that the first parent allele increased WA% at both loci. One QTL for °BS was detected on chromosome 8 (LOD 4.96; PVE 15.9%; Add 0.14). Three QTLs for YM% were identified on chromosomes 7, 8, and 14 with LOD scores of 3.02, 4.48, and 3.15 (PVE 8.2%, 12.6%, and 8.8%), with additive effects of 0.67, 0.84, and -0.69, indicating that Daepung allele increased YM% on chromosomes 7 and 8, while Galchae parent allele increased it on chromosome 14. Two QTLs for TR% were detected on chromosomes 8 and 13 (LOD 4.12 and 2.61; PVE 12.1% and 7.5%), with additive effects of -0.98 and 0.78. Notably, QTL regions on chromosome 8 for WA% and YM%, as well as for °BS and TR%, co-localized at the same marker positions, suggesting potential pleiotropic effects (Table 3, Fig. 4). Overall, the additive effects observed in both RIL populations indicate that alleles from both parents can contribute positively or negatively to the traits, providing valuable information for future marker-assisted selection.
Our results are consistent with Döttinger et al. (2025), who reported several medium to large effect QTL for tofu quality traits, including pleiotropic loci on chromosomes 4 and 10 and major effect QTL on chromosomes 5 and 13. Similarly, we identified QTL with moderate effects and co-localizations for WA% and TR% (chromosomes 1 and 6) and for WA%, YM%, °BS, and TR% (chromosome 8), suggesting pleiotropic regulation. In both studies, favorable alleles were derived from different parents, reinforcing the importance of diverse germplasm for enhancing tofu processing quality. Notably, none of the QTL associated with WA% overlapped with previously reported loci for related traits such as seed weight (Döttinger et al. 2025, Kim et al. 2010, Kurasch et al. 2018, Orf et al. 1999, Teng et al. 2008, Yan et al. 2014). In contrast, we observed QTL for WA% that were proximal to loci for soaking-related traits on chromosomes 1, 6, and 20, though positioned at slightly different genetic distances (cM) (Kurasch et al. 2018). Furthermore, recent studies identified QTL for soymilk weight (SM), TW, TY, TH, and tofu value (TV) (Döttinger et al. 2025, Kurasch et al. 2018), Among these, two loci for TH and TV were detected near TR% associated QTL on chromosomes 8 and 13, but again at distinct genetic positions (cM). A recent study highlighted that hard beans markedly reduce SM, TY, and TV, and identified two QTL on chromosomes 1 and 14 (Döttinger et al. 2025). Notably, one of these loci was located near our QTL for YM% on chromosome 14, although positioned at a slightly different genetic distance (cM). Collectively, these results highlight tofu quality as a complex trait shaped by multiple pleiotropic QTL. Their consistent detection across studies underscores their breeding relevance and potential for integration into marker-assisted selection to develop soybean cultivars with improved tofu processing quality.
The absence of significant QTLs for tofu yield in this study does not necessarily indicate a lack of genetic control but rather reflects the genetic and biological complexity of the trait. (Kurasch et al. 2018). Tofu yield is likely governed by numerous minor-effect loci, and limited allelic diversity between the parental lines may have reduced detectable variation. In addition, the precision and reproducibility of phenotyping, which is highly sensitive to processing conditions, could have limited QTL detection. To address these challenges, approaches such as genome-wide association studies (GWAS) and eQTL mapping offer greater power and resolution.
In Silico Identification of Tofu Quality-related Genes
In the Daepung×Saedanbaek population, three QTLs for WA% were identified on Chromosomes 1, 6, and 20, with intervals ranging from 117 to 2,028 kbp and harboring 17-208 candidate genes. A QTL for °BS was detected on Chromosome 17 (409 kbp, 39 candidates). For TR%, three QTLs were mapped to Chromosomes 1 and 6, collectively spanning 340-2,028 kbp and harboring 365 candidates. Additionally, a QTL for YT% was identified on Chromosome 3 (1,021 kbp, 27 candidates). The number of candidate genes for each QTL is summarized in Table 2.
In the Daepung×Galchae population, QTLs for WA% were detected on Chromosomes 8 and 15, spanning 208 and 581 kbp, respectively, with 98 candidate genes identified in total. A QTL for °BS was mapped to Chromosome 8 within a 564 kbp interval containing 73 candidate genes. YM% was associated with three QTLs on Chromosomes 7, 8, and 14, spanning 102-488 kbp and comprising 78 candidate genes. Two QTLs for TR% were identified on Chromosomes 8 and 13, spanning 187-563 kbp and harboring 96 candidate genes. The number of candidate genes for each QTL is presented in Table 3. Based on QTL mapping and functional annotation in the Daepung×Saedanbaek population (Supplementary Tables 5, 6, 7), Glyma.06g173500 (Chromosome 6) within a QTL for WA% and TR%, and Glyma.03g132700 (Chromosome 3) within a QTL for YT%, were identified as candidate genes involved in 1,3-β-D-glucan (callose) biosynthesis and degradation, respectively (Table 4). Callose, a key component of the seed coat, regulates cell wall rigidity and permeability. Its hydrolysis by 1,3-β-D-glucanase enables rapid modulation of callose content, affecting cell wall hardness (Chen & Kim 2009, Chen et al. 2024).
In legumes, a thick callose layer, often accompanied by suberin or wax, contributes to seed coat impermeability (Leubner 2003). Supporting this, introduction of a gene inducing hard seed coat in the soybean cultivar Tachinagaha enhanced seed coat integrity and reduced permeability, accompanied by accumulation of 1,4-β-glucan in palisade cells (Jang et al. 2015). These observations suggest that callose content modulates WA, which is a critical determinant of tofu yield. Indeed, prolonged soybean soaking increased tofu yield from 45.9% to 55.4%, independent of raw seed protein content, highlighting the influence of cultivar-specific physicochemical traits (Park et al. 1985, Sim et al. 2020). In addition, a recent study identified a QTL containing the candidate gene Gy4, encoding the glycinin G4 subunit on chromosome 10, associated with TW and exhibiting pleiotropic effects on TH and TV (Döttinger et al. 2025). Consistently, in the present study, we identified Glyma.06g173500 within a QTL region associated with WA%, which also displayed pleiotropy with TR%.
Taken together, these findings suggest that the identified candidate genes may play central roles in regulating seed coat properties and WA%, thereby directly influencing TY. Functional validation of Glyma.06g173500 and Glyma.03g132700 will be essential to elucidate the underlying biochemical mechanisms and to develop molecular markers for breeding soybean cultivars with enhanced tofu quality.
Conclusion
This study aimed to develop simplified evaluation methods for soybean processing traits and to identify QTLs and candidate genes associated with tofu quality using two RIL populations. Two RIL populations, Daepung×Saedanbaek (81 lines) and Daepung×Galchae (122 lines), exhibited substantial variation in tofu quality traits. These patterns indicate both populations are suitable for QTL analysis. Correlation analysis indicated strong positive relationships between WA% and TR%, as well as YM% and YT%, with some differences between populations. QTL mapping identified eight QTLs in each population: in Daepung×Saedanbaek, for WA% (Chr. 1, 6, 20), °BS (Chr. 17), TR% (Chr. 1, 6), and YT% (Chr. 3); in Daepung×Galchae, for WA% (Chr. 8, 15), °BS (Chr. 8), YM% (Chr. 7, 8, 14), and TR% (Chr. 8, 13). Candidate genes Glyma.06g173500 (1,3-β-D-glucan biosynthesis) and Glyma.03g132700 (1,3-β-D-glucanase) were associated with WA% and YT%, likely through regulation of seed coat permeability.
Overall, the study provides an efficient method for large-scale evaluation of tofu quality traits and identifies candidate genes that could serve as molecular markers for breeding superior soybean cultivars.
Acknowledgement
This research was supported by a research grant from the Rural Development Administration project (Project No. RS-2024-00398383), Republic of Korea.
Supplementary Data
Supplementary Tables 1-7 is available on the journal website.
Distribution for tofu quality traits in Daepung (P1)×Saedanbeak (P2) RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. (E) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soymilk yield, TR%: tofu residue, YT%: tofu yield. The red curve denotes normal distribution of the trait and green curve represents the estimated best-fit normal density.
Fig. 2
Distribution for tofu quality traits in Daepung (P1)×Galchae (P2) RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. (E) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soymilk yield, TR%: tofu residue, YT%: tofu yield. The red curve denotes normal distribution of the trait and green curve represents the estimated best-fit normal density.
Fig. 3
Identification of QTLs in Daepung×Saedanbeak RIL population. (A) WA%. (B) °BS. (C) TR%. (D) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soy milk yield, TR%: tofu residue, YT%: tofu yield.
Fig. 4
Identification of QTLs in Daepung×Galchae RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. WA%: water absorption, °BS: Brix of soymilk, YM%: soy milk yield, TR%: tofu residue, YT%: tofu yield.
Table 1
Correlation coefficients about tofu quality traits of in two RIL populations.
zTrait: WA%; water absorption, °BS; Brix of soymilk, YM%; soymilk yield, TR%; tofu residue, YT%; tofu yield.
References
1. Chen J, Xu X, Liu W, Feng Z, Chen Q, Zhou Y, Sun M, Gan L, Zhou T, Xuan Y. 2024. Plasmodesmata function and callose deposition in plant disease defense. Plants 13: 2242
6. Hou Y, Huang L, Xing G, Yuan X, Zhang X, Dai D, Yuan X, Chen X, Xue C. 2025. Structure and functional properties of proteins from different soybean varieties as affected by the 11S/7S globulin ratio. Foods 14: 755
7. Jang SJ, Sato M, Sato K, Jitsuyama Y, Fujino K, Mori H, Takahashi R, Benitez ER, Liu B, Yamada T, Abe J. 2015. A single-nucleotide polymorphism in an endo-1,4-β-glucanase gene controls seed coat permeability in soybean. PLoS One 10: 1-19.
9. Jung J, Kim JM, Lee T, Heo J, Shin I, Lee JS, Kang S. Development of fast and simple processing method to evaluate tofu traits in soybean breeding system. Korean J Breed Sci 54: 34-42.
11. Kim G, Park A, Kim WJ, Moon CY, Kang BH, Kim SH, Choi YM, Ha BK. 2021. Development of a simple enzymatic method for screening sucrose content in legume seeds. Plant Breed Biotechnol 9: 250-258.
12. Kim HK, Kim YC, Kim ST, Son BG, Choi YW, Kang JS, Park YH, Cho YS, Choi IS. 2010. Analysis of quantitative trait loci (QTLs) for seed size and fatty acid composition using recombinant inbred lines in soybean. J Life Sci 20: 1186-1192.
13. Kim HT, Ko JM, Baek IY, Jeon MK, Han WY, Park KY, Lee BW, Lee YH, Jung CS, Oh KW, Ha TJ, Moon JK, Yun HT, Lee JH, Choi JK, Jung JH, Lee SS, Jang YJ, Son CK, Kang DS. 2014. Soybean cultivar for tofu, 'Saedanbaek' with disease resistance, and high protein content. Korean J Breed Sci 46: 295-301.
14. Kim JM. 2024. Genetic linkage map of K-SoyNAM population. figshare. Dataset. doi:10.6084/m9.figshare.27926514.v1.
15. Kim S-L, Berhow MA, Kim JT, Chi HY, Lee SJ, Chung I-M. 2006. Evaluation of soyasaponin, isoflavone, protein, lipid, and free sugar accumulation in developing soybean seeds. J Agric Food Chem 54: 10003-10010.
16. Kinsella J. 1979. Functional properties of soy protein. J Am Oil Chem Soc 56: 242-258.
17. Kurasch AK, Leiser WL, Bachteler K, Miersch M, Hahn V, Würschum T. 2018. Evaluation of the genetic architecture of tofu traits in soybean towards genomics-assisted breeding. Plant Breed 137: 873-882.
18. Lee YG, Jeong N, Kim JH, Lee K, Kim KH, Pirani A, Ha BK, Kang ST, Park BS, Moon JK, Kim N, Jeong SC. 2015. Development, validation and genetic analysis of a large soybean SNP genotyping array. Plant J 81: 625-636.
19. Leubner G. 2003. Functions and regulation of β-1,3-glucanases during seed germination, dormancy release and after-ripening. Seed Science Research 13: 17-34.
21. Meng L, Li H, Zhang L, Wang J. 2015. QTL IciMapping: Integrated software for genetic linkage map construction and quantitative trait locus mapping in biparental populations. The Crop Journal 3: 269-283.
22. Modgil R, Tanwar B, Goyal A, Kumar V. Tanwar B, Goyal A, 2021. Soybean (Glycine max). editors. Oilseeds: Health attributes and food applications. Springer Singapore. Singapore: pp. 1-46.
23. Oh HS, Lee KH, Yoon S. 1981. Preparation of soymilk yogurt and the related studies. Korean J Nutr 14: 175-181.
24. Oh Y, Cho S, Kim K, Kim Y, Kim T, Kim J, Paik K, Back I, Ran W, Kim H. 2009. A new soybean cultivar, "Galchae" for sprout with brown seed coat, small seed size and high sprout yielding. Korean J Breed Sci 41: 324-327.
25. Orf J, Chase K, Jarvik T, Mansur L, Cregan PB, Adler F, Lark K. 1999. Genetics of soybean agronomic traits: I. Comparison of three related recombinant inbred populations. Crop Sci 39: 1642-1651.
26. Park KY, Moon JK, Yun HT, Lee YH, Kim SL, Ryu YH, Kim YH, Ku JH, Roh JH, Lee ES, Ha KS, Kim IJ, Son CK, Kim SK, Kim SD, Moon HP. 2005. A new soybean cultivar for fermented soyfood and tofu with high yield, "Daepung". Korean J Breed Sci 37: 111-112.
27. Park MR, Seo MJ, Lee YY, Park CH. 2016. Selection of useful germplasm based on the variation analysis of growth and seed quality of soybean germplasms grown at two different latitudes. Plant Breed Biotechnol 4: 462-474.
28. Park YG, Park BD, Choi KS. 1985. Changes in ultrastructure of tissues, characteristics of protein and soybean curd yield with increased soaking time during soybean curd processing. J Korean Soc Food Nutr 14: 381-386.
29. Patel J, Patel S, Cook L, Fallen BD, Koebernick J. 2025. Soybean genome wide association study of seed weight, protein, and oil content in the southeastern USA. Mol Genet Genomics 300: 43
32. Qin P, Wang T, Luo Y. 2022. A review on plant-based proteins from soybean: Health benefits and soy product development. Journal of Agriculture and Food Research 7: 100265.
33. Sanjukta S, Rai AK. 2016. Production of bioactive peptides during soybean fermentation and their potential health benefits. Trends in Food Science & Technology 50: 1-10.
35. Sim EY, Lee C, Park HY, Lee Y, Choi HS, Lee S, Kim HS, Kang BK, Chun A, Kim M, Kwak J, Jeon YH. 2020. Quality characteristics of tofu made from Korean soybean cultivars. Food Eng Prog 24: 54-61.
36. Teng W, Han Y, Du Y, Sun D, Zhang Z, Qiu L, Sun G, Li W. 2008. QTL analyses of seed weight during the development of soybean (Glycine max L. Merr.). Heredity 102: 372-380.
37. Yan L, Li YH, Yang CY, Ren SX, Chang RZ, Zhang MC, Qiu LJ. 2014. Identification and validation of an over-dominant QTL controlling soybean seed weight using populations derived from Glycine max×Glycine soja. Plant Breed 133: 632-637.
Fig. 1
Distribution for tofu quality traits in Daepung (P1)×Saedanbeak (P2) RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. (E) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soymilk yield, TR%: tofu residue, YT%: tofu yield. The red curve denotes normal distribution of the trait and green curve represents the estimated best-fit normal density.
Fig. 2
Distribution for tofu quality traits in Daepung (P1)×Galchae (P2) RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. (E) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soymilk yield, TR%: tofu residue, YT%: tofu yield. The red curve denotes normal distribution of the trait and green curve represents the estimated best-fit normal density.
Fig. 3
Identification of QTLs in Daepung×Saedanbeak RIL population. (A) WA%. (B) °BS. (C) TR%. (D) YT%. WA%: water absorption, °BS: Brix of soymilk, YM%: soy milk yield, TR%: tofu residue, YT%: tofu yield.
Fig. 4
Identification of QTLs in Daepung×Galchae RIL population. (A) WA%. (B) °BS. (C) YM%. (D) TR%. WA%: water absorption, °BS: Brix of soymilk, YM%: soy milk yield, TR%: tofu residue, YT%: tofu yield.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
콩에서 두부특성 연관 양적형질유전자좌(QTLs) 탐색
Correlation coefficients about tofu quality traits of in two RIL populations.
