Detection of QTLs for seedling characteristics in barley (Hordeum vulgare L.) grown under hydroponic culture condition

Background Seedling characteristics play significant roles in the growth and development of barley (Hordeum vulgare L.), including stable stand establishment, water and nutrients uptake, biotic resistance and abiotic stresses, and can influence yield and quality. However, the genetic mechanisms underlying seedling characteristics in barley are largely unknown and little research has been done. In the present work, 21 seedling-related characteristics are assessed in a barley double haploid (DH) population, grown under hydroponic conditions. Of them, leaf age (LAG), shoot height (SH), maximum root length (MRL), main root number (MRN) and seedling fresh weight (SFW) were investigated at the 13th, 20th, 27th, and 34th day after germination. The objectives were to identify quantitative trait loci (QTLs) underlying these seedling characteristics using a high-density linkage map and to reveal the QTL expression pattern by comparing the QTLs among four different seedling growth stages. Results A total of 70 QTLs were distributed over all chromosomes except 4H, and, individually, accounted for 5.01%–77.78% of phenotypic variation. Out of the 70 detected QTLs, 23 showed a major effect on 14 seedling-related characteristics. Ten co-localized chromosomal regions on 2H (five regions), 3H (two regions) and 7H (three regions) involved 39 QTLs (55.71%), each simultaneously influenced more than one trait. Meanwhile, 9 co-localized genomic regions involving 22 QTLs for five seedling characteristics (LAG, SH, MRL, MRN and SFW) at the 13th, 20th, 27th and 34th day-old seedling were common for two or more growth stages of seedling. QTL in the vicinity of Vrs1 locus on chromosome 2H with the favorable alleles from Huadamai 6 was found to have the largest main effects on multiple seedling-related traits. Conclusions Six QTL cluster regions associated with 16 seedling-related characteristics were observed on chromosome 2H, 3H and 7H. The majority of the 29 regions identified for five seedling characteristics were selectively expressed at different developmental stages. The genetic effects of 9 consecutive expression regions displayed different developmental influences at different developmental stages. These findings enhanced our understanding of a genetic basis underlying seedling characteristics in barley. Some QTLs detected here could be used for marker-assisted selection (MAS) in barley breeding. Electronic supplementary material The online version of this article (10.1186/s12863-017-0562-y) contains supplementary material, which is available to authorized users.


Background
As the earliest domesticated crop and the fourth most important cereal crop in the world, barley (Hordeum vulgare L.) is not only widely used as human food and animal feed, but is also an ideal model of genetic research because it is early maturing, diploid, self-fertilizing and has a short growth period [1][2][3]. However, its growth and production are also greatly affected by seedling characteristics. The seedling stage (from seedling emergence to jointing) is considered as the most critical stage for barley production, and plays significant role in barley growth and development, including stable stand establishment, water and nutrients uptake, biotic resistance and abiotic stresses, and can influence yield and quality [4,5]. In fact, strong seedling vigor and rapid seedling growth are important breeding targets in barley, as well as in other crops [6][7][8]. Seedling vigor has been identified to be associated with salt tolerance [9], drought tolerance [10,11], nutrient uptake [12], weed competitiveness [8] and yield [13,14]. It is difficult to assess seedling characteristics because a majority of important seedling characteristics are complex, quantitative traits determined by an array of developmental processes, genetic and environmental factors. The genetic mechanisms controlling seedling characteristics in barley are poorly known and little research has been done in this regard. Genetic variation for seedling characteristics in barley varieties does exist, and dissecting the genetic and molecular basis of seedling characteristics is necessary for genetic improvement of barley cultivars to enhance seedling vigor.
The information on the genetic basis underlying seedling characteristics, especially for root characteristics, is limited because it is difficult to obtain reliable phenotypic data in a large number of seedlings and in a complex external environment. It is also difficult to continuously measure these traits from the same seedling due to destruction of the seedlings in field experiments [15][16][17]. Therefore, an alternative approach is to examine seedling characteristics under controlled conditions using hydroponic culture, which has several advantages over field and other conventional techniques, including: (i) ease of investigation of seedling characteristics (as compared with field experiment), (ii) exclusion of soil and environment interference to increase measurement repeatability, (iii) investigation of large numbers of lines in short period of time, and (iv) precise control of the concentration of nutrient concentration [18]. At present, this method has been widely used in rice [19][20][21], maize [15,22,23] and wheat [16,24,25], and many QTLs have been detected at various water and nutrient regimes. Some researchers have used this method to identify QTLs associated with salt tolerance [3,26], waterlogging tolerance [27][28][29] and nitrogen stress tolerance [30] in barley.
With the development of molecular markers and continuous encryption of genetic map, QTL analysis has become a powerful tool to dissect complex traits and identify chromosomal regions harboring genes that control these quantitative traits [31,32], and has been extensively used for genetic dissection of important traits in barley [33][34][35]. The QTLs underlying yield [34,36,37] and quality [35,38,39] traits at later growth stages have been characterized previously, and the QTLs for seedling characteristics reported were mainly identified under salt tolerance [29,40,41], waterlogging tolerance [29,42], drought tolerance [43] and nitrogen stress tolerance [30], while seedling characteristics related to grain yield and quality at early developmental stages are not well investigated.
The objectives of our study were to identify QTLs underlying seedling characteristics in hydroponics at seedling growth stages, and compare its QTL expression patterns among four different seedling growth stages.

Evaluation of seedling characteristics
The DH population and two parents were grown hydroponically in a greenhouse as shown in Fig. 1a. Figure 1b showed phenotypic difference in the seedling of the Huadama 6 and Huaai 11. Huadama 6 and Huaai 11 have significant differences in seedling characteristics. All traits values, except FOLW, in Huadamai 6 were higher than those in Huaai 11. The T-test showed that two parents were significantly different on all traits (p < 0.05) but FILW and FOLW ( Table 1).
The 21 seedling characteristics in DH population were also significantly different, with a highly phenotypic variation. The variation coefficient ranged from 6.88%-35.29%. Heritability ranged from 89.86% to 97.05% (Table 1). Frequency distribution histogram of the 21 seedling characteristics were shown in Additional file 1: Figure S1. Shapiro-Wilk test showed that some of the characteristics fitted the normal distribution, including LAG (S2), SH (S1-S4), MRL (S1), SFW (S1), LL (except FILL), LW and LA. In addition, transgressive segregation was observed for all 21 seedling characteristics.

Detection of QTLs
A total of 70 QTLs on all barley chromosomes (except 4H) were detected in 17 of the 21 seedling characteristics using ICIM mapping (  Fig. 2). Phenotypic variation explained by single QTL ranged from 6.03% to 73.86% for LAG, from 5.49% to 24.90% for SH, from 7.85% to 32.39% for SFW, from 8.08% to 16.23% for MRL and from 7.71% to 31.28% for MRN. Interestingly, we found that only one QTL for SH was detected at all four stages, 8 QTLs for LAG, SH, SFW and MRL were detected for two or three stages, but also some QTLs were specific for either S1, S2, S3 and S4 (Table 4; Fig. 4). In addition, we also observed that some QTLs, such as the QTL at 123 cM on chromosome 2H, not only expressed at all stages for SH but also affected the SFW at the S2 stage. Similar QTLs were identified at 127/129 cM and 137/139 cM on chromosome 2H and at 30/32 cM and 80/81 cM on chromosome 3H. Two major QTLs for LAG (qLAG2-228) and SFW (qSFW2-202) explaining more than 20% of the phenotypic variation at more than one stage of seedling growth was identified on chromosome 2H. We analyzed the temporal-dynamics contribution of QTLs detected at any of the four stages to the proportion of explained phenotypic variation for five seedling characteristics (Fig. 5). The results showed that the proportion of explained phenotypic variance of some QTLs remained on a stable level at all stages of seedling growth; whereas, others substantially changed among the four stages of seedling growth.

Discussion
Seedling characteristics include stem-and-leaf and root systems. A good stem-and-leaf system in the seedling stage is the basis for seedlings growth potential of robust, strong plants that are resistant to stress. To sustain a strong seedling vigor and rapid seedling growth, seedling must produce abundant nutrition. The amount of nutrition production mainly depends on the first few leaves of the seedling [44]. In addition, growth and development of healthy plants mainly rely on the strong root system (long MRL, high RDW and enough MRN), which were closely related to final grain yield and quality [45]. Some studies indicated that MRL and MRN were one of the most desirable traits contributing to drought avoidance, and RDW was another important parameter  for the roots associated with water use efficiency and long-term drought [46,47]. Recent studies have also confirmed that most of the seedling traits, such as leaf area (LA), seedling height (SH), seedling fresh weight (FW), seedling dry weight (DW), shoot dry weight (SDW) and root dry weight (RDW), were closely and positively correlated with yield component traits [48]. Previous genetic research on barley were focused mainly on the later stage; thus, the genetic mechanism underlying seedling characteristics is still lacking. In this study, we assessed QTLs underlying 21 seedling characteristics (including five seedling characteristics assessed at four stages of seedling growth) under hydroponic culture conditions and compared the QTL expression pattern among four different seedling growth stages.

Advantages of hydroponic culture
For larger populations, it is difficult to assess seedling characteristics, especially whole-root systems, and for multi-stage barley planted in soil. Hydroponics culture allowed us to assess the same seedling over multi-stages. Genes or QTLs expression under hydroponic conditions might be different from field trials. However, the expression of genes or QTLs is always affected by the environment, while complex environments and uneven fertilizing in the field trial will affect the genetic mechanism of seedling characteristics, and may not reflect the intrinsic genetic program of barley seedling characteristics [19]. Thus, stable and consistent environmental control is needed to assess seedling characteristics. Recent studies have shown that QTLs identified under hydroponics conditions corresponded to QTLs detected in field trials, and suggested that a hydroponic system is a fast and cost-effective method for early QTL detection and marker-assisted allelic selection [30]. Consequently, we established a hydroponic system with homogeneous-growth conditions to evaluate barley-seedling characteristics. Temperature, humidity and lighting conditions were controlled by the automatic control system, and the nutrient solution was renewed every seventh day to ensure adequate nutrition. QTLs identified in this environment could reflect the intrinsic genetic mechanisms underlying barley seedling characteristics.

QTL for seedling characteristics
With the same DH population used in this study, Ren et al. [49][50][51][52], Liu et al. [33] and Wang et al. [34] identified certain QTLs for physiological and morphological traits of flag leaf and some agronomic and quality traits (Additional file 3: Figure S3). Similar to their study, most of the QTLs identified in this study were also located on chromosome 2H and 7H. We found that some QTLs for seedling characteristics were co-located with QTLs for yield or yield-related traits identified in previous studies. For example, a significant QTL qSH2-191 for SH detected in all stages of seedling growth was close to the SNP marker 2_514353957, which is likely the same to the yield-  Table 6. S1 to S4 represented the 13th, 20th, 27th and 34th day after germination, respectively related QTL qSms2-7 and qTgw2-1 reported by Wang et al. [34] (Table 2; Additional file 3: Figure S3). In addition, the co-localized QTLs for THLL, SELA and THLA were detected on chromosome 7H and are likely the same to qRWC7-9 and qSPD7-9 for physiological and morphological traits of flag leaf at the pre-filling stage reported by Liu et al. [33] (Table 2; Additional file 3: Figure S3). These results suggested that seedling characteristics and yield are related to a certain extent.
In previous studies, QTLs conferring seedling height (SH) [29,30], root dry weight (RDW) [46,[53][54][55] and stem-and-leaf dry weight (SLDW) [30,42,46,53,55] were reported on seven linkage groups, while QTLs for seedling fresh weight (SFW), leaf age (LAG) and main root number (MRN) were rarely reported in barley. QTLs for root length were identified on chromosome 1H, 2H, 3H, 4H, 5H and 7H [46,[53][54][55], and QTLs conferring root fresh weight (RFW) were previously reported on chromosome 1H, 2H and 5H [29,42]. Recently, Hoffmann et al. [30] analyzed QTL underlying number of leaves longer than 2 cm after 14 days and length of the youngest completely unfolded leaf after 14 days, and found six QTLs for leaf number on chromosome 2H, 4H, 6H and 7H, and nine QTLs for leaf length on all seven chromosomes in wild barley introgression lines (S42ILs). By using GrainGenes3.0 website (http://wheat.pw.usda.gov/GG3/) to compare the genetic markers with those results present in this study, we found that some QTLs identified in our study are likely same to the QTLs reported in the past research. Such as, qSH3-44 for SH is likely the same to the QTL QHei.S42IL-3H reported by Hoffmann et al. [30]. Arifuzzaman et al. [46] reported that QTL QSdw.S42.2H.c for SLDW on chromosome 2H was  Table 1. c QTL-likelihood curves of LOD scores showed the locations of QTL for some seedling characteristics on chromosome 2H in the vicinity of Vrs1 locus. Length of the rectangles corresponded to the 1.5-LOD support intervals based on the results of ICIM. Vrs1 locus was indicated by hatched bars. The black dash line indicated that the LOD value was 3.0 linked with the marker bPb-8143, and since the bPb-8143 is near the morphological marker Vrs1, as inferred from GrainGenes3.0 (http://wheat.pw.usda.gov/ GG3/), we suggested that a co-localized QTLs for SLDW, SFW, SLFW, MRN, FILA, SELW, SELA and FOLA identified in our study is likely the same to the QTL QSdw.S42.2H.c. This QTL may be an important locus for controlling stem-and-leaf related traits in barley. The SNP marker 2HL_34260490 may be useful for marker-assisted selection (MAS) in barley breeding.

Selective expression of QTLs
According to developmental genetics, different QTLs may have different expression dynamics during trait development [7,57]. Previous studies focused mainly on late-growth stages, where analysis was limited to the performance of a trait at a fixed time or stage of ontogenesis [33,34,58]. In the present study, we assessed five seedling characteristics (LAG, SH, SFW, MRL and MRN) at four stages of seedling growth (13th, 20th, 27th and 34th day-after germination) to reveal the QTL expression pattern.
A total of 42 QTLs involved 29 regions were detected for these five traits at four stages of seedling growth  Table 6. S1 to S4 represented the 13th, 20th, 27th and 34th day after germination, respectively Fig. 5 The temporal contributions of QTLs detected at any of the four stages to the proportion of explained phenotypic variation for five seedling characteristics. Trait abbreviations refer to Table 6. S1 to S4 represented the 13th, 20th, 27th and 34th day after germination, respectively (Table 2; Fig. 4). Although the number of QTL expressed in each stage was approximately the same, only 9 regions were continually expressed in two or more different stages of seedling growth (Table 4; Figs. 3b and 4). Of these, only one of the nine regions was persistently expressed at all stages (Table 4; Figs. 4 and 5). The majority of the 29 regions for these five seedling characteristics were selectively expressed at different developmental stages. This result is in agreement with other studies where different QTLs could be identified at different growth stages, where only a few co-localized QTLs were detected at all stages [7,59,60]. The regions that continually expressed at different growth stages might play an important role in the growth and development of these characteristics. However, these regions do not always produce the same effect at each stage but show different developmental influences in different seedling growth stages. For example, two co-localized regions for SH were identified on chromosome 2H and 3H; its contribution to the phenotypic variance undergoes temporal changes. On chromosome 2H, a colocalized region contribution to the phenotypic variance decreased from 18.61% at S1 to 17.74% at S2 and then gradually increased until S4 (22.05%). Differently, a colocalization region on chromosome 3H contribution to the phenotypic variance gradually decreased from 13.49% at S1 to 10.92% at S3 and disappeared at S4. Besides, the co-localized regions for MRL, SFW and LAG contribution to the phenotypic variance also showed similar temporal changes. Generally, our results revealed that the genetic architecture of the five seedling characteristics shows dynamic temporal changes during growth and development.

Pleiotropy or linkage of QTLs
We found here that many QTLs controlling multiple seedling characteristics were located at the same or overlapping marker interval on some chromosomes. This implied an existence of pleiotropic QTL or tightly linked QTL in our study. Xu et al. [7] suggested that if two QTL peaks are located very close to each other, and the 1-LOD support intervals completely or mostly overlapped, these two QTLs would be regarded as a single QTL having pleiotropic effect. In the present study, we delimited 1.5-LOD support intervals to confirm whether these two QTLs controlling multiple seedling characteristics can be identified as the pleiotropic QTL or tightly linked QTL. Based on this assumption, we found ten regions on chromosomes 2H, 3H and 7H that showed pleiotropic or tightly linkage effects on different seedling characteristics (Table 3).
Five regions on chromosome 2H were observed to simultaneously affect multiple seedling characteristics. Of these, a significant region with pleiotropic or tightly linkage effects on SFW, SLFW, SLDW, MRN, SELW, FILA, SELA and FOLA was close to both the SNP marker, 2HL_34260490 and the morphological marker, Vrs1 ( Fig. 3C; Table 3). The Vrs1 locus (controlling row type) has a pleiotropic effect on multiple traits [34,[61][62][63]. For example, Marquez-Cedillo et al. [61] and Wang et al. [34] reported that the Vrs1 locus had a pleiotropic effect on thousand grain weight, spike length, plant height, spikelet number on main spike and grain number per plant. Furthermore, some malting quality traits, such as grain-protein-percentage, diastatic power and soluble/ total protein ratio, were also coincident with the Vrs1 locus reported by Marquez-Cedillo et al. [62]. Sato et al. [63] suggested that the fusarium head blight (FHB) resistance QTL in Vrs1 locus reported in two-row × six-row crosses might be pleiotropic effect of Vrs1 locus. Our study also showed a possible pleiotropic effect of the Vrs1 locus and provided evidence to further support previous suggestions.
On chromosome 3H and 7H, we also found that some regions showed pleiotropic effects on different seedling characteristics (Table 3). A pleiotropic QTL or tightly linked QTL, detected between SNP marker 3HL_29995337 and 3HL_32354397 near the SSR marker Bmag13, was related to SH and SELA. Examining the 3H linkage maps of Xue et al. [64] and GrainGenes3.0 (http://wheat.pw.usda.gov/GG3/) revealed that the marker bPb-6504 is near to the marker Bmag13, indicating that this QTL is likely the same to the qSPL3c associated with spikes per line. In the previous studies using the same DH population, QTL for heading date (Qhd3-13), culm length (Qcl3-13), two internode length(Qitw3-13) and three internode length (Qith3-13), as reported by Ren et al. [49,51], and QTLs for stomatal conductance (qGs3-13) and transpiration rate (qTr3-13) reported by Liu et al. [33] were also near this region (Additional file 3: Figure S3). A QTL located at 93/94 cM on chromosome 7H showed that pleiotropic or tightly linkage effects on THLL, SELA and THLA is likely the same to the QTL (qRWC7-9) for relative water content and QTL (qSPD7-9) for relative chlorophyll content, as reported by Liu et al. [33]. These results provided favorable support for pleiotropic QTL or tightly linked QTL identified in our study. Further, this pleiotropic QTL or tightly linked QTL may be used for marker-assisted selection (MAS) and genetic improvement of barley seedling characteristics in barley breeding.
The phenomenon of pleiotropic effect or linked for QTLs universally exists in various crop genomes. Pleiotropic effects of genetic loci are thought to play a critical role in evolution, reflecting functional and developmental relationships among phenotypes [65]. The application of linked QTL and pleiotropic QTL for breeding purposes should be carried out with care, especially if the pleiotropy and linkage are in negative direction [7]. For example, the Huadamai 6 alleles with the pleiotropic QTL or tightly linked QTL on chromosome 3H increased main roots number but decreased maximum root length (Table 3). However, a pleiotropic QTL with opposite effects on the different traits complicated its application in barley improvement.

QTL clusters in the genome
In our study, we found that most of QTLs were located on chromosomes 2H, 3H and 7H (Figs. 2 and 3a). It is noteworthy that many QTLs gathered in the same chromosome regions, even for the unrelated traits (Table 5). For instance, one QTL cluster region for RDW, SLDW, MRL, SFW and RFW was mapped on chromosome 2H close to the marker GBM1218. Liu et al. [33] reported QTL clusters for physiological and morphological traits of flag leaf (net photosynthetic rate, stomatal conductance, flag leaf area, flag leaf length, flag leaf width, relative chlorophyll content and leaf nitrogen concentration) in the similar region. Similarly, Ren et al. [49] detected QTLs for heading date traits in this region. Wang et al. [34] also detected QTLs for spikelet number on main spike in this region. Considering all information here, we suggested that this region may be a credible region for cluster of QTLs.
The most important QTL cluster was detected on chromosome 2H close to the morphological marker Vrs1, which was clustered with 22 QTLs (31.43%) for LAG, SH, MRL, MRN, SFW, SLFW, SLDW, RFW, RDW, FOLL, SELW, FILA, SELA and FOLA with contribution of 12.05%-32.39% (Table 5; Fig. 2). Wang et al. [34] reported QTL clusters for some agronomic traits (spikelet number on main spike, grain number per spike, spikelet number per plant, grain number per plant, grain weight per plant and thousand grain weight) in this chromosome region. Hori et al. [66] detected QTLs for plant height, spike exsertion length and thousand-kernel weight in this region. In addition, QTLs for spike number, floret number, grain number and hundred grain weight were also detected in this region by Chutimanitsakun et al. [67]. QTLs for seedling traits in this region may be a result of the pleiotropic effect of the Vrs1 locus or expression of the closely linked genes for Vrs1.
Another significant QTL cluster underlying LAG, FILL, FILA and SELA was detected on chromosome 2H between SNP marker 2HL_18227895 and 2_598509820 (Table 5; Fig. 2). The effects from Huadamai 6 were positive for LAG, but negative for FILL, FILA and SELA. Castro et al. [68] detected QTLs for photoperiod response in this region. Wang et al. [34] also detected QTLs for length of main spike, spike density, thousand grain weight and grain weight per spike in this region, and suggested that the SNP marker 2_598509820 in this region could be used for marker-assisted selection. In addition, another three QTL clusters were obtained on chromosome 3H and 7H, all of which enhanced development of stem-and-leaf traits (Table 5; Fig. 2).
QTL clustering in barley was repeatedly reported in some studies [33,34,56,69]. In our study, we demonstrated several significant QTL clusters of seedling characteristics in barley under hydroponic culture. The genetic mechanism for this widespread phenomenon may be a linkage of genes and of pleiotropic effect of a single QTL in the same genomic region [70,71]. However, whether it is linkage or pleiotropy, further research using fine mapping and cloning of QTLs or genes are needed validation.

Conclusions
In this research, a number of genomic regions containing many seedling-related QTLs involved in different traits and diverse seedling growth stages were detected. The two parents used for mapping population construction are different in row types (six-rowed dwarfing barley cultivar Huaai 11 and two-rowed barley cultivar Huadamai 6). We found that chromosome 2H close to the morphological marker Vrs1 contained the most important QTL cluster regions controlling multiple seedling characteristics. Some QTLs detected in this research could be used as a potential target for marker-assisted selection (MAS) in barley breeding.

Plant materials and hydroponic culture experiments
The barley double haploid (DH) population used in the current study was derived by anther culture from a cross  [33,34,[49][50][51][52]. In the present study, 122 lines and the two parental cultivars were evaluated at seedling stage using the hydroponic culture (Fig. 1a) Each seedling was anchored to holes 1.5 cm diameter in foam board with sterilized sponge. Each foam board had 5 rows of holes, and each row had 6 holes for a total of 30 seedlings per board. The distance between two seedlings was 3.6 cm (between rows) and 4.0 cm (within a row), respectively. Each tank contained eight liters of revised Hoagland's nutrient solution [72], which was renewed every seventh day to prevent nutrient exhaustion. When renewed the nutrient solution, each tank moves randomly to ensure that same environmental conditions. The solution PH was adjusted to 6.5 using diluted NaOH and HCl before refreshing. The two parents and DH lines were grown in completely randomized design with six replicates. In each replication, one seedling was used for data collection.

Phenotyping of seedling characteristics
The phenotypes measured included leaf age (LAG), seedling height (SH, cm), maximum root length (MRL, cm), main roots number (MRN) and seedling fresh weight (SWF, g) at four growth stages (13th, 20th, 27th and 34th days after germination) on the same six seedlings for each line. Root fresh weight (RFW, g), stemand-leaf fresh weight (SLFW, g), root dry weight (RDW, g) and stem-and-leaf dry weight (SLDW, g) were only measured at the 34th day after germination. For convenience, we used S1, S2, S3 and S4 to represent the 13th, 20th, 27th and 34th day after germination, respectively. In addition, leaf length (LL, cm), leaf width (LW, cm) and leaf area (LA, cm 2 ) were measured in four leaves according to the leaf order of seedling. LL was measured for first leaf length (FILL, cm), second leaf length (SELL, cm), third leaf length (THLL, cm) and fourth leaf length (FOLL, cm). LW was measured for first leaf width (FILW, cm), second leaf width (SELW, cm), third leaf width (THLW, cm) and fourth leaf width (FOLW, cm). LA as measured for first leaf area (FILA, cm 2 ), second leaf area (SELA, cm 2 ), third leaf area (THLA, cm 2 ) and fourth leaf area (FOLA, cm 2 ). The methods of measurement are listed in Table 6.

Statistical analysis
The mean phenotypic values of the 21 seedling characteristics obtained from hydroponic culture were subjected to The roll-leaf length measurement standard is, if the length of roll-leaf less than 1/3 the length of next leaf was recorded as 0.1, between 1/3 and 1/2 the length of next leaf was recorded as 0.3, between 1/2 and 3/4 the length of next leaf was recorded as 0.5, between 3/4 and the length of next leaf was recorded as 0.7, longer than the length of next leaf was recorded as 0. statistical analysis. Normality of distribution was tested using the method of Shapiro-Wilk. The descriptive statistics analyses were performed using SPSS programs (IBM SPSS Statistics, Chicago, IL, USA, http://www.ibm.com/ analytics/us/en/technology/spss). The heritability was estimated [73]. Frequency distribution and QTL-likelihood maps for the seedling characteristics were drawn using the Origin programs (OriginLab, Northampton, MA, USA, http://www.originlab.com). P-value higher than 0.05 was deemed as significance level.

QTL analysis
The high-density genetic linkage map constructed by Ren et al. [50] for the 'Huaai 11 × Huadamai 6' population was used in the QTL analysis, which contains 1894 SNP markers and 68 SSR markers, covering all 7 chromosomes and spanning 1375.80 cM of the barley genome with an average marker distance of 0.7 cM. The estimates of QTL positions and effects were determined by inclusive composite interval mapping (ICIM) [74] using the software QTL IciMapping 4.1 [75]. The mapping method was selected ICIM-ADD (additive effects) in 'BIP (QTL mapping in biparental populations)' function to analyze each trait from a single environment. Significant LOD (likelihood-of-odd) threshold for declaring a QTL of each trait was determined by 1000 permutations test with a Type 1 error of 0.05 [76]. The scanning step size was set at 1.0 centimorgan (cM) and the probability in stepwise regression (PIN) was 0.001. The rule recommended by Liu et al. [33] was used to name the QTLs, which were mapped on linkage groups using the software MapChart 2.2 [77]. The QTLs whose percentage of phenotypic variation explained (PVE) exceeded 20%, were considered as major QTL, and otherwise minor QTL. If the 1.5-LOD support interval of two QTLs overlapped, these two QTL would be defined as co-localized QTLs. However, we independently counted each QTL to clearly specify the number of detected QTLs. Graingenes website (http://wheat.pw.usda.gov/GG3/) was used to compare the maker information.

Additional files
Additional file 1: Figure S1. Frequency distribution of 21 seedling characteristics in DH population. P value of Shapiro-Wilk test for each stage was shown, the hypothesis of normal distribution was accepted when P > 0.05 (significant at P = 0.05), and the trend lines of the accepted normal distribution were shown. Trait abbreviations refer to the Table 6. S1 to S4 represented the 13th, 20th, 27th and 34th day after germination, respectively. (JPEG 5396 kb) Additional file 2: Figure S2. QTL likelihood map for 17 seedling characteristics in the DH population grown under hydroponic culture conditions using inclusive composite interval mapping. Genetic maps (all chromosomes together) of barley linkage groups were shown in the abscissa and LOD scores of each trait in the ordinate. The significant LOD threshold was determined to be 3.0 by 1000 permutations test. Trait abbreviations refer to the Table 6. S1 to S4 represented the 13th, 20th, 27th and 34th day after germination, respectively. (JPEG 4454 kb) Additional file 3: Figure S3. QTL locations for previous studies and current studies detected in the Huaai 11 × Huadamai 6 DH population. QTLs were mapped to chromosomes using peak position and nearest marker. (