Profiling of seed fatty acid composition in 1025 Chinese soybean accessions from diverse ecoregions

Ahmed M. Abdelghny, Shengrui Zhng, Muhmmd Azm,Abdulwhb S. Shibu, Yue Feng, Ynfei Li, Yu Tin, Huilong Hong,Bin Li,, Junming Sun,

aThe National Engineering Laboratory for Crop Molecular Breeding, MARA Key Laboratory of Soybean Biology (Beijing), Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,Beijing 100081,China

bCrop Science Department, Faculty of Agriculture, Damanhour University, Damanhour 22516,Egypt

AB S T R A C T The stability of soybean(Glycine max L.Merrill)oil is determined mainly by its fatty acid(FA)composition.We evaluated the FA composition of 1025 Chinese soybean accessions collected from diverse ecoregions and grown in multiple locations and years. We observed highly significant differences (P ＜0.001) between accessions in palmitic acid (PA), stearic acid (SA),oleic acid (OA), linoleic acid (LA), and linolenic acid (LNA) contents. Growth year affected(P ＜0.001)the abundance of all FAs except PA. The mean PA, SA, OA, LA, and LNA contents were 12.2%,3.8%,21.5%,54.2%,and 8.3%,respectively.The geographical origin of the accession influenced seed FA composition,indicating that accessions originating in each ecoregion tend to have distinct FA composition. We observed significant positive correlations among the three locations and between the two years, suggesting the high heritability and stability of individual accessions across contrasting environments. We also observed a relatively high negative correlation between the contents of OA and both LA and LNA (r = −0.90 and −0.59,respectively, each significant at P ＜0.001), providing a potential entry point for developing strains producing oil with higher OA and lower LA and LNA levels.These would be appropriate for specialized use in the food industry. Our results will be useful in breeding soybean with improved quality to meet human nutritional and industrial needs.

1. Introduction

Soybean (Glycine max L. Merrill) is a major source of protein and oil [1]. Protein is the most prominent component of soybean seeds, accounting for 35%-55% (dry weight), and soy protein has long been consumed in foods such as tofu,soymilk, tempeh, natto, and soybean sprouts [2]. Oil components account for approximately 18%-20% of dry seed weight,and soybean oil is also widely used for human consumption,as well as in a broad range of industrial applications[3,4].

Soybean seed oil comprises five fatty acids (FAs): the saturated FAs palmitic acid (PA) and stearic acid(SA) and the unsaturated FAs oleic acid (OA), linoleic acid (LA), and linolenic acid (LNA). Several studies have indicated that temperature is an environmental factor affecting FA composition, as it strongly influences the biosynthesis of FAs,especially unsaturated FAs [5-7]. FA composition also varies greatly among growing regions [8,9]. This variation could be due to any of a number of environmental factors that are highly variable across locations [10]. For example, variations in seed FA profile by location and environment may be attributable to climatic differences, as temperature can critically influence FA biosynthesis in plants during the ripening period [11,12] or, alternatively, to past breeding to meet different market needs and consumer preferences [13].OA levels tend to decrease from high-to low-latitude regions,representing a decline with rising temperature, in contrast to LA and LNA levels [14-16]. Saturated FAs are less affected by temperature than unsaturated FAs[5].Sunshine duration and rainfall also influence soybean seed FA composition, as sunshine duration correlates positively with PA, SA,and LNA levels and negatively with OA level, whereas the effect of rainfall on these FAs is opposite[17].

Chinese soybean germplasm represents the primary germplasm resource for the species and harbors rich phenotypic and genetic diversity[18].Distributed across a wide ecological range, this soybean germplasm harbors variation in multiple nutrient components as a result of adaptation to complex climate and soil environments[19]as well as to human needs[20]. The primary germplasm grown in regions of China is likely to carry sufficient genetic variation to withstand many challenges encountered in plant breeding[18].These remarkable soybean genetic resources,which include local landraces,cultivars, and wild soybeans, have played a central role in improving soybean yield and quality through different breeding programs[21].A collection of 28,580 soybean accessions is currently conserved in the Chinese Gene Bank [22]. A core collection of 2794 accessions was established from this collection that represents 11.8% of the full collection and 73.6% of its genetic diversity [23]. Our study used 1025 accessions from this core collection, accounting for 36.7% of the core collection.

The three ecoregions of China in which the soybean accessions investigated in this study originate, namely the Northern Region,Huanghuaihai Region,and Southern Region,span the area from 23°N to 51°N latitude[24,25]with variable environmental conditions such as temperature, rainfall, and sunshine duration.These regions represent the main growing areas of Chinese soybean germplasm [26]. Previous studies[9,17,27] investigated the quality traits of individual accessions when grown within their region of origin.In the present study, we investigated the seed FA composition of a large collection of Chinese germplasm with diverse ecoregion origins grown in three environments, Beijing, Anhui, and Hainan,in two years.We expected the study of this large and diverse germplasm collection to assist in the selection of accessions that maintain high levels of desired FAs when grown outside their native ecoregions, and thus to enhance future breeding programs aimed at soybean quality improvement.The objective of this study was to investigate the effect of geographical origin on seed FA composition in a large panel of diverse Chinese soybean accessions with the goal of guiding soybean quality improvement for FA composition.

2. Materials and methods

2.1. Plant materials

A total of 1025 Chinese soybean accessions from three ecoregions, the Northern Region, Huanghuaihai Region, and Southern Region, provided by the soybean genetic resource research group of the Institute of Crop Sciences, CAAS, were used. They included 925 landraces and 100 modern cultivars originally collected across China and selected from the Chinese primary core collection, which was developed to capture as much as possible of the phenotypic diversity and geographic distribution of the collection of cultivated soybean conserved in the Chinese National Soybean Gene Bank(CNSGB) [21,23]. The number and type of accessions and geographical data for each origin are shown in Table S1.

2.2. Field experiments

Field trials were conducted in 2017 and 2018 at Changping(40°13′N, 116°12′E), Beijing and Sanya (18°24′N, 109°5′E),Hainan province, and in 2017 at Hefei (33°61′N 117°0′E),Anhui province. All accessions were planted at Changping on June 12, 2017 and June 14, 2018; at Sanya on Nov. 14, 2017 and Nov. 16, 2018; and at Hefei on June 5, 2017. Soil pH and total nitrogen, phosphorus, and potassium levels were respectively 8.22, 80.5 mg kg−1, 68.7 mg kg−1, and 12.31 g kg−1at Changping [28]; 6.6, 35.91 mg kg−1, 56 mg kg−1, and 134.40 g kg−1at Hefei; and 5.27, 98.59 mg kg−1, 39.68 mg kg−1,and 80.78 g kg−1at Sanya. The field experiments were laid out in a randomized incomplete block design, with the three planting locations used as replications. The mean monthly temperature, rainfall, and sunshine of the three experimental sites are shown in Table S2. At each location, seeds of each accession were planted in a 3-m row, spaced 0.5 m apart between rows and 0.1 m between plants within rows. After emergence, they were thinned to leave only uniform healthy plants. Plots were fertilized with 15 t ha−1organic fertilizer containing 30 kg N ha−1, 60 kg P ha−1, and 50 kg K ha−1during field preparation before sowing. Weeds were controlled by postemergence application of 2.55 L ha−1of acetochlor, as well as hand weeding during the growing season. When plants reached physiological maturity, plots were harvested manually.

2.3. FA extraction and determination

The five essential FAs (palmitic, stearic, oleic, linoleic, and linolenic) were derivatized to their methyl esters and their abundances determined by gas chromatography [29]. Briefly,20 g of seeds from each accession were finely ground with a Sample Preparation Mill (Retsch ZM100, Rheinische, Germany). Then, 300 mg of powder from each sample was weighed out using an analytic balance (Sartorius BS124S,Gottingen, Germany) and transferred to a 2-mL centrifuge tube preloaded with 1.0 mL n-hexane. This mixture was held for 20 min at 65 °C and shaken for 10 s every 5 min. Next,1.0 mL sodium methoxide solution was added to the mixture and it was shaken for 10 min on a twist mixer (TM-300, AS ONE,Osaka, Japan)at 65°C to allow full methyl esterification of the FAs, followed by centrifugation at 12,000 ×g for 2 min.The supernatant was assayed to determine the concentrations of the methyl esters of the five FAs using a GC-2010 gas chromatograph (Shimadzu Inc., Kyoto, Japan) with flame ionization detector. The chromatographic separation was performed on an RTX-WAX column (Restek, Germany, 30 m length×0.25 mm internal diameter×0.25 mm thickness)and the temperature gradient was as follows: initially, the temperature was set at 180 °C for 1.5 min, then increased to 210 °C at a rate of 10 °C min−1, held at 210 °C for 2 min,increased to 220 °C at a rate of 5 °C min−1, and held at 220 °C for 5 min. The carrier gas was nitrogen, at a flow rate of 54 mL min−1, and 1 μL of each sample was injected.Area was normalized to quantify the five FA concentrations using a GC2010 workstation[29].

2.4. Geographical distribution mapping

Geographical distribution maps of soybean seed FA composition were constructed with ArcGIS 10.0 (ESRI, Redlands, CA,USA, http://desktop.arcgis.com/en/arcmap/) using ordinary kriging interpolation [30]. ArcGIS is widely employed in geographic information system (GIS) applications, as it can be readily used to create maps, compile geographic data and manage spatial data. To process the maps in our study, we subjected the five FA means across locations and years, as well as geographical factors(longitude and latitude),to kriging interpolation,which assigns weights to known sample points to estimate the values of unknown sample points.The kriging interpolation formula can be expressed as where z(x)is the value of the unknown sample point,ziis the value of the ith known sample point near the unknown sample point,n is the number of known sample points,and wiis the weight of the ith known sample point applied to the unknown sample point.

2.5. Data analysis

Analysis of variance (ANOVA) for seed FA composition was performed using PROC GLM in SAS 9.2 [31]. Differences were considered significant when the P value was ＜0.05 (P ＜0.05).Multiple comparisons of means were performed using Tukey’s honestly significant difference test. Boxplots were drawn to show the distribution and variation of seed FA composition in the three ecoregions. The five FA means across locations and years were subjected to Pearson’s correlation analysis and principal component analysis (PCA)among the three ecoregions. PCA was performed to show the contribution of each FA composition to the total variation among different regions and accession types.Figures showing boxplots, correlations,and PCA statistics were created with R 3.5.0[32].

3. Results

3.1.Variation in seed FA composition in 1025 Chinese soybean accessions

Overall,levels of PA and SA ranged from 10.2% to 15.2% and from 2.6% to 7.5%, with means of 12.2% and 3.8%, respectively. The coefficients of variation(CVs)of PA and SA were 5.9% and 11.4%,respectively (Table 1). OA levels ranged from 13.3% to 36.1%,with an average of 21.5% and a CV of 15.7%(Table 1).LA and LNA ranged from 40.4% to 63.9% and 3.9% to 12.8% with averages of 54.2% and 8.3% and CVs of 5.3% and 14.6%,respectively(Table 1).The accession with the highest PA was ZDD09581(15.2%),which originated in the Huanghuaihai Region, whereas that with the lowest PA (10.2%) was ZDD10812, which also originated in the Huanghuaihai Region. For SA, ZDD03901, originating in the Huanghuaihai Region, had the highest level (7.5%), whereas ZDD05905, originating in the Southern Region, had the lowest(2.6%). For OA, ZDD17157, originating in the Southern Region,had the highest level(36.1%),whereas ZDD11085,originating in the Huanghuaihai Region,had the lowest level(13.3%).For both LA and LNA,accessions originating in the Southern Region had both the highest levels(ZDD17256 with 63.9%LA and ZDD06322 with 12.8%LNA)and the lowest levels(ZDD04373 with 40.4%LA and ZDD17157 with 3.9% LNA). Additional information about accessions with desirable levels of FA levels is shown in Table S3.

3.2.Accession type affects soybean seed FA composition

We observed significant differences (P ＜0.001) in seed FA composition between landraces and cultivars(Fig.1,Table S4).Levels of all FAs differed significantly (P ＜0.01) by cultivation year, except PA, which showed a non-significant difference(Table S4). Interaction between accession type and year did not significantly influenced the levels of all FAs (Table S4).The cultivars tested showed higher levels of PA, SA, and OA(12.4%, 4.0%, and 23.7%, respectively) than the landraces(12.2%, 3.8%, and 21.2%, respectively) (Fig. 1A-C), whereas landraces showed higher levels of both LA and LNA (54.4% and 8.4%, respectively) than cultivars (52.4% and 7.5%,respectively) (Fig. 1D, E).

3.3.Correlations between FA contents

The correlations between soybean seed components are described in Fig. 2. There was a significant negative correlation between OA and LA (r = −0.90, P ＜0.001, denoted***), as well as between OA and LNA levels(r=−0.59***),whereas there was a significant positive correlation between LA and LNA(r=0.29***).The PA content was significantly negatively correlated with both OA and LA contents, but not with LNA content. SA content showed a significant negative correlation with LA and LNA content, but a significant positive correlation with OA content.

Table 1-Variation in FA composition(%)among soybean grown in three ecoregions and two years(2017 and 2018).

3.4. Principal component analysis (PCA) based on ecoregion and accession type

The results of PCA revealed that the first two components(PC1 and PC2) accounted for 68.1% of the total variation observed. Two PCAs were fitted, grouping accessions based on ecoregion (Fig. S1-A) and accession type (Fig. S1-B). In both biplots,the first component(PC1),which accounted for 45.3% of the total variation, revealed that the highest contribution to total variation was from OA content(41.6%), followed by SA content (34.7%), whereas PA (0.1%)made the smallest contribution. PC2 accounted for 22.8% of the total variation,and PA contributed most to the variation of PC2 (57.9%), followed by SA (37.8%), whereas LNA contributed the least (0.15%). The biplot for ecoregion (Fig.S1-A)showed that variation across the three ecoregions was well captured in this analysis. The majority of accessions from the Southern Region were distributed around the LA and LNA content variables, indicating the tendency of the Southern Region to contain accessions with higher levels of LA and LNA. Accessions from the Northern Region were closest to the PA, SA, and OA content variables, reflecting a tendency of that ecoregion toward higher levels of these FAs. The biplot based on accession type (cultivar and landrace) (Fig. S1-B) showed that most cultivars were distributed close to the PA, SA, and OA content variables,implying that cultivars retain high contents of these components. However, the dispersion of landraces around the LA and LNA variables showed that they tended to have higher levels of these FAs.

Fig.1-Comparison of seed FA composition between soybean cultivars and landraces.(A)Palmitic acid.(B)Stearic acid.(C)Oleic acid.(D)Linoleic acid.(E)Linolenic acid.Different lowercase letters(a and b) indicate statistically significant differences at P ＜0.05.

3.5.Ecoregion and geographical factors affect soybean seed FA composition

ANOVA revealed highly significant differences (P ＜0.001)among ecoregions for all FAs(Table S4).Interaction of year×ecoregion showed a significant effect(P ＜0.05)on SA and OA contents and a highly significant effect (P ＜0.01) on LA content (Table S4). In contrast, PA and LNA contents were not significantly affected by this interaction(Table S4).The variation in FA composition among the ecoregions is shown in Table 1 and Fig. 3A-E. Accessions from the Northern Region contained the highest levels of PA(12.4%),SA(3.9%),and OA (23.4%), followed by the Huanghuaihai Region accessions, with PA, SA, and OA levels of 12.4%, 3.9%, and 21.2%, respectively. Accessions from the Southern Region showed the lowest contents of these FAs (12.1%, 3.7%, and 21%, respectively) (Fig. 3A-C). There were no significant differences between accessions from the Northern and Huanghuaihai Regions for PA level, and between the Huanghuaihai and Southern Regions for OA level. In contrast,both LA(Fig.3D)and LNA(Fig.3E)contents showed a trend of increase from high(north)to low latitude(south).Accessions from the Southern Region had the highest levels of both LA (54.5%) and LNA (8.8%), followed by accessions from the Huanghuaihai Region, with LA level of 54.3% and LNA level of 8.2%, whereas accessions from the Northern Region showed the lowest levels of LA (52.8%) and LNA(7.4%). There was no significant difference in LA level between accessions from the Northern and Huanghuaihai Regions.

Fig.2- Correlations between soybean seed contents of five FAs.Significant differences are indicated by*P ＜0.05,**P ＜0.01,***P ＜0.001;values without asterisks were not significant at P ＜0.05.Values are Pearson's correlation coefficients (r). PA,palmitic acid;SA,stearic acid;OA, oleic acid;LA,linoleic acid;LNA, linolenic acid.

Fig.3- Distribution of seed FA composition in soybeans collected from three ecoregions.(A)Palmitic acid.(B)Stearic acid.(C)Oleic acid.(D)Linoleic acid.(E) Linolenic acid.NR,Northern Region;HR, Huanghuaihai Region;SR,Southern Region.Different lowercase letters(a,b,and c)indicate statistically significant differences at the P ＜0.05 level.

The geographical distribution of seed FA composition based on ecoregion origin is shown in Fig. 4A-E. The correlations between seed FA composition and the geographical factors of latitude,longitude,and altitude are presented in Table 2. The average FA composition of soybean accessions across locations and years correlated with the latitude,longitude, and altitude of their corresponding ecoregion origins. PA level showed a significant positive correlation with latitude, but no relationship with longitude or altitude.SA level showed significant positive correlations with both latitude and longitude, but not altitude. LNA level showed a significant negative correlation with latitude and longitude but not with altitude. By contrast, OA and LA levels were influenced by all three geographical factors, but in opposite directions: OA level showed a positive correlation with latitude and longitude but a negative correlation with altitude,whereas LA level correlated negatively with latitude and longitude but positively with altitude.

Fig.4-Geographical distribution of soybean seed FA composition.(A)Palmitic acid.(B)Stearic acid.(C)Oleic acid.(D)Linoleic acid.(E)Linolenic acid.Based on 1025 accessions from regions of China.

The geographical distribution maps (Fig. 4C, D) showed opposite trends in OA and LA levels,across China,and were in agreement with the correlations between FA contents and geographical factors. In particular, provinces along the eastern coast of China represent hotspots for accessions with higher OA and lower LA levels.These areas are characterized as lowland areas belonging to the third tier of altitude(altitude ＜500 m), at relatively high latitude and longitude.In contrast, the accessions in the highland region that includes the Yunnan-Guizhou and Loess plateaus, which belong to the second altitude tier(1000 m ＜altitude ＜2000 m)and are also relatively lower in latitude and longitude,showed an opposite profile of higher LA and lower OA levels.

Correlation coefficients among the three planting locations were also significant and positive for all the seed FA compositions (Table S5). Similarly, correlations between the two study years were significant and positive for all seed FAs,with pairwise coefficients of 0.59***,0.39***,0.64***,0.67***,and 0.64***for PA,SA,OA,LA, and LNA,respectively.

Table 2-Correlations between soybean FA compositions and geographical traits of the studied ecoregions.a

4. Discussion

The significant differences observed in FA composition among cultivars, accession types, and ecoregions confirmed the importance of genetic factors in breeding soybeans for desired FA profiles. Recent studies [33,34] have revealed similar findings. The variation in seed FA composition between landraces and cultivars in our study was in agreement with previous studies on other crops including lupine(Lupinus albus L.) [35], Brassica napus and B. rapa [36], and common bean (Phaseolus vulgaris L.) [37]. Soybean landraces are initially adapted to contrasting environments, making them suitable,to some extent,for local production.Cultivated soybeans, however, are artificially selected for desired traits,especially oil composition. As cultivated soybeans have higher oil content and OA level than wild soybean, oil composition can be used as a reference index for evolutionary classification in soybean [38]. Our results confirmed that landraces had higher levels of LA and LNA (Fig. 1D, E) than cultivars, and cultivars had higher levels of OA (Fig. 1C) than landraces,suggesting that FA composition may be associated with soybean domestication.

We observed a wider range of LA and LNA levels (40.4%-63.9% and 3.9%-12.8%,respectively)than was seen in previous studies (38.9%-58.9% and 5.1%-11.7%, respectively) [17,27],emphasizing the broad genetic variation in the panel of accessions we used. More strikingly, the overall average contents of PA (12.2%), LA (54.2%), and LNA (8.3%) in this study were higher than those observed in the previous studies(11.8%,50.2%,and 7.7%, respectively).

The positive correlation between PA and SA contents revealed in this study favors the prospect of increasing the contents of both FAs simultaneously through selection to improve the stability of soybean oil.The negative correlations between SA and both LA and LNA contents may be attributable to the influence of temperature on unsaturated FA contents in diverse environments,in agreement with findings of Bellaloui et al. [16]. This correlation provides double benefits, as a higher level of SA is desirable in baking fats like margarine and shortenings [39], whereas minimal levels of LA and LNA are favored for avoiding oil instability[40].The negative correlation of OA with both LA and LNA, which was also reported in other studies [14,41], may have been caused by the effect of elevated temperatures during seed fill,resulting in an accumulation of OA and decline of LA and LNA contents. This correlation should make it possible to produce oils with higher OA levels, which positively affect human health by reducing low-density lipoprotein cholesterol[40,42,43], as well as increasing oil stability against oxidation. Increasing LNA consumption is thought [44] to be beneficial to cardiovascular health. Our PCA findings identified OA as the most discriminative variable for PC1 among the five FAs [41,45]. Thus, OA level can be used to identify accessions with targeted levels of multiple FAs.The PCA findings and the correlations among FA compositions in this study suggest that direct or indirect selection of accessions containing high levels of beneficial FAs is an achievable goal.

Previous studies have shown that accessions of diverse geographical origins have diverse seed FA composition,owing not only to highly variable environmental factors such as temperature, rainfall, and light but also to geographical factors such as latitude and longitude [9,12,17]. Song et al.[17] investigated climatic and geographical effects on the FA composition of soybean accessions of different origins and established that the latitude in which the accessions were grown was irrelevant to their levels of PA, OA, LA, and LNA,but that climatic factors significantly influenced FA levels,emphasizing the discrepant influence of geographical and environmental factors on seed composition. In contrast to their findings that only SA content showed a positive association with latitude (declining from north to south), we observed significant positive associations of PA, SA, and OA levels with latitude. This positive correlation means that levels of PA, SA, and OA increase with higher latitude and lower temperature, which characterize conditions in the Northern Region of China, providing evidence for the superiority of accessions from the Northern Region in producing higher levels of PA,SA,and OA.In contrast,LA and LNA levels showed significant negative correlations with latitude, increasing from north to south, indicating a superiority of accessions from the Southern Region with respect to these components.More generally,our findings imply the existence of alleles conferring adaptability in Chinese soybean accessions and suggest that these accessions can be cultivated in diverse locations irrespective of ecoregion (geographical origin).These findings support an inference that some quality traits, such as FA composition, were domesticated and improved through selection and accordingly show regional differentiation patterns due to an increasing trend in frequency of specific alleles from lower latitude (southern regions) to higher latitude (northern regions), and vice versa[46].Together,these results suggest that breeders should pay more attention to the regional origin of accessions when developing improved strains of soybean [47]. They also encourage the exploitation of these valuable genetic resources in breeding programs.

The significant positive correlations among the three locations and between the two years for FA composition indicate that many accessions show relatively stable increasing or decreasing trends across different environments. Also, the non-significant interactions of both year × cultivar and year ×accession type suggest the adaptability of these accessions when cultivated under diverse environmental and geographical conditions. This adaptability can be further explained by the presence of alleles that promote adaptability.

Our geographical distribution maps were constructed using ArcGIS software, which provided a spatial distribution pattern of the five FA compositions over three Chinese ecoregions important in soybean production. Many studies have used ArcGIS-based maps in order to obtain clear view of the phenomena concerned [48-52]. The use of ArcGIS to construct the geographical maps facilitated our identification of the dominance of certain ecoregions in producing strains with specific FA compositions and further showed the trends in the contents of each FA across China.Because China has an extensive collection of soybean accessions under cultivation,spanning a wide diversity and incorporated into complex cropping systems, it has become a priority to investigate trends in soybean seed quality in various ecoregions using a large number of varieties.The selection of candidate varieties is indispensable for breeding soybeans with improved seed quality. Thus, the variation we identified among the three ecoregions should make it possible to select adaptable accessions possessing higher average contents of particular FAs in diverse locations. For example, accessions in the Northern Region shared high levels of PA, SA, and OA,suggesting that such accessions will also produce high levels of these FAs even when grown in other regions with different climatic and geographical conditions.Similarly,the collection of accessions in the Southern Region constitute a useful collection with relatively high levels of LA and LNA under various environmental and geographic conditions. With respect to accession-type-based selection,some landraces and cultivars with high levels of each of the five FAs shown in Table S3 provid a vital resource for soybean breeders.Collectively,our study suggests some accessions such as ZDD09581,ZDD03901,ZDD17157, ZDD17256, and ZDD17157 as prominent landraces with highly desirable levels of PA, SA, OA, LA, and LNA,respectively. These promising accessions could be used by soybean breeders in developing further improved varieties with higher contents of specific desired FAs.

5. Conclusions

The diverse origins of soybean accessions influence soybean seed FA composition, revealing a tendency for accessions from each ecoregion to have distinct FA composition. The correlations between geographical factors and seed composition in this study confirmed previously reported trends in seed FA contents in accessions derived from the three ecoregions across diverse growth environments. The results of correlation coefficients among soybean seed FA components that we obtained should help guide breeders in selection of accessions with preferable profiles. Our findings suggest that the process of domestication and recent breeding programs extensively influenced seed FA composition,resulting in a marked improvement of seed FA composition in promising modern cultivars. In general, the various geographical origins can provide different collections of accessions with desired levels of seed FA composition to meet the nutritional demands of a wide range of consumers with diverse dietary needs.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2019.11.002.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

Acknowledgments

We thank Dr.Lijuan Qiu from Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,for providing all of the soybean accessions and for her helpful suggestions on the manuscript.We also thank Dr. Xiaobo Wang from Anhui Agricultural University for helping in growing soybean accessions in Anhui province. This study was supported by Ministry of Science and Technology of China (2016YFD0100201, 2016YFD0100504 and 2017YFD0101400),National Science and Technology Major Project of China(2016ZX08004-003),National Natural Science Foundation of China (No. 31671716), Beijing Science and Technology Plan Project (Z16110000916005), and CAAS (Chinese Academy of Agricultural Sciences) Agricultural Science and Technology Innovation Program.