Molecular characterization of stable QTL and putative candidate genes for grain zinc and iron concentrations in two related wheat populations

Major QTL for grain zinc and iron concentrations were identified on the long arm of chromosomes 2D and 6D. Gene-based KASP markers were developed for putative candidate genes TaIPK1-2D and TaNAS10-6D. Micronutrient malnutrition is one of the most common public health problems in the world. Biofortification, the most attractive and sustainable solution to surmount malnutrition requires the development of micronutrient enriched new crop cultivars. In this study, two recombinant inbred line (RIL) populations, ZM175/XY60 and ZM175/LX987, were used to identify QTL for grain zinc concentration (GZnC), grain iron concentration (GFeC) and thousand grain weight (TGW). Eight QTL for GZnC, six QTL for GFeC and five QTL for TGW were detected. Three QTL on chromosomes 2DL and 4BS and chromosome 6A showed pleiotropic effects on all three traits. The 4BS and 6A QTL also increased plant height and might be Rht-B1a and Rht25a, respectively. The 2DL locus within a suppressed recombination region was identified in both RIL populations and the favorable allele simultaneously increasing GZnC, GFeC and TGW was contributed by XY60 and LX987. A QTL on chromosome 6DL associated only with GZnC was detected in ZM175/XY60 and was validated in JD8/AK58 RILs using kompetitive allele-specific PCR (KASP) marker K_AX-110119937. Both the 2DL and 6DL QTL were new loci for GZnC. Based on gene annotations, sequence variations and expression profiles, the phytic acid biosynthesis gene TaIPK1-2D and nicotianamine synthase gene TaNAS10-6D were predicted as candidate genes. Their gene-based KASP markers were developed and validated in a cultivar panel of 343 wheat accessions. This study investigated the genetic basis of GZnC and GFeC and provided valuable candidate genes and markers for breeding Zn- and Fe-enriched wheat.


Introduction
Grain yield is the most important trait in breeding, and the so-called Green Revolution contributed greatly to crop production (Gupta et al. 2020a;Welch and Graham 1999).However, due to emphasis solely on yield, modern productive cereal cultivars have displaced traditional crops that had higher micronutrient contents (Welch 2002;Welch and Graham 1999).As a result, micronutrient malnutrition, especially regarding deficiencies of zinc (Zn) and iron (Fe), has become a global health issue, currently afflicting more than 700 million children and women (FAO 2022).Current strategies for improving human nutrition mainly include dietary diversification, pharmaceutical supplementation, industrial fortification and biofortification, as reviewed in Ali and Borrill (2020).Due to the limited diet for people in poverty-stricken areas and insufficient investment in industrial fortification, biofortification is the most promising, cost-effective and sustainable strategy to address malnutrition.
Bread wheat (Triticum aestivum L.) is one of the most important staple crops, providing about 50% of the daily energy intake in developing countries, and its nutrition quality has a direct impact on human health (Cakmak 2007;Cakmak et al. 2010;Ortiz-Monasterio et al. 2007).To meet the normal Zn and Fe requirements of human body, the concentrations of Zn and Fe in whole grains should be 37 mg/kg and 59 mg/kg, respectively (Bouis et al. 2011), while the existing wheat cultivars in China have GZnC and GFeC of 30 mg/kg and 45 mg/kg, respectively (Liu et al. 2014a, b).To develop cultivars with high mineral accumulation, it is necessary to exploit genetic diversity for grain Zn concentration (GZnC) and grain Fe concentration (GFeC) in wheat and its wild relatives (Monasterio and Graham 2000;Ortiz-Monasterio et al. 2007;Velu et al. 2014).In recent years, diverse bi-parental and cultivar wheat populations have been used to detect QTL associated with GZnC and GFeC (Gupta et al. 2020b;Tong et al. 2020), and several GZnC and GFeC hotspots in wheat genomic regions have been identified using meta-QTL analysis (Shariatipour et al. 2021a, b).Thus far, only Triticum dicoccoides-derived gene Gpc-B1 affecting grain protein content, GZnC and GFeC has been cloned and utilized in some breeding programs (Distelfeld et al. 2007; Uauy et al. 2006).
Negative correlations between grain yield and mineral concentrations, also known as "dilution effect", are frequently observed in wheat (Fan et al. 2008;Murphy et al. 2008).The widely deployed dwarfing genes Rht-B1b and Rht-D1b led to a significant increase in wheat grain yield, but reduced grain micronutrient concentrations to some extent (Jobson et al. 2018;Velu et al. 2017).Thus, it is important to improve yield and nutrition traits simultaneously in breeding.To achieve this objective, breeders can either pyramid multiple favorable genes for different traits by marker-assisted selection or use QTL or genes with pleiotropic effects on both yield and nutrition (Hao et al. 2014).
In this study, two recombinant inbred line (RIL) populations with Zhongmai 175 (ZM175) as a common parent, and Xiaoyan 60 (XY60) and Lunxuan 987 (LX987) as second parents, were studied.QTL analyses were made of GZnC, GFeC and thousand grain weight (TGW) using high-density genetic maps and grain from multi-environmental trials.We aimed to: (a) identify genetic loci associated with high GZnC, GFeC and TGW; (b) validate the major QTL in different genetic backgrounds using breeder-friendly kompetitive allele-specific (KASP) markers; and (c) predict putative candidate genes.

Plant materials and field trials
RIL populations were generated from crosses ZM175/XY60 (250 F 7 lines), and ZM175/LX987 (146 F 5 lines).ZM175 released in 2007, was a leading cultivar in the Northern Winter Wheat Region (He et al. 2015).XY60 was derived from Xiaoyan 6, an important founder parent in Chinese wheat breeding (Li et al. 2008).LX987 released in 2003, was developed by recurrent selection in a population segregating for male sterility (Li et al. 2019).The ZM175/XY60 population was used in previous studies on morphological, physiological and agronomic traits (Luo et al. 2020(Luo et al. , 2021)).
The ZM175/XY60 population was grown in five environments; Gaoyi (37°33′N, 114°26′E) and Gaocheng (37°27′N, 113°30′E) in the 2018-2019 cropping season, and Gaoyi, Gaocheng and Beijing (39°56′N, 116°20′E) in 2019-2020.The ZM175/LX987 population was grown in four environments, Gaoyi, Gaocheng, Beijing and Urumqi (42°45′N, 86°37′E) in 2019-2020.For QTL validation, a RIL population derived from a cross of Jingdong 8/Aikang 58 (JD8/ AK58, 254 F 6 lines) was grown in nine environments including Gaoyi, Gaocheng and Beijing in 2016-2017, 2017-2018and 2018-2019(Wang et al. 2021)).A panel of 343 wheat cultivars/lines, mainly from the Northern Winter Wheat region and Yellow and Huai River valleys was grown in six environments, including Gaoyi, Gaocheng and Beijing in 2019-2020 and 2020-2021.Each population was grown in randomized complete block design with two replicates for each environment.Each line was planted in a single 1 m row with 0.2-0.3m spacing between rows, and approximately 40 grains were sown in each row.To minimize the heterogeneity of soil Zn, about Page 3 of 16 217 25 kg/ha of ZnSO 4 .7H 2 O was applied prior to planting the Gaoyi and Gaocheng trials in each year.As green manure was routinely returned to field in Beijing and only one crop per year was grown at Urumqi additional Zn fertilizer was considered unnecessary.

Phenotypic evaluation and statistical analysis
Plots were harvested at physiological maturity when grains were completely dried.Grain samples were hand threshed and carefully cleaned to remove debris and foreign matter.About 15 g samples were taken for determination of GZnC and GFeC, using a "bench-top," non-destructive, energydispersive X-ray fluorescence spectrometry (EDXRF) instrument (model X-Supreme 8000, Oxford Instruments Plc, Abingdon, UK).This instrument had been standardized for high-throughput screening of GZnC and GFeC in whole grain wheat (Paltridge et al. 2012).About 500 intact grains were to estimate TGW using a SC-G Seed Test System (WSeen Testing Technology Co., Ltd., Hangzhou, China).Plant height was measured during the grain-filling stage from the ground to spike tip (excluding awns).
Analysis of variance (ANOVA), correlation analysis and t-tests were performed using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA).The best linear unbiased estimation (BLUE) for phenotypic data across environments was calculated using the ANOVA function in QTL IciMapping v4.1 software (Li et al. 2007).Broad sense heritability (H) was calculated following the formula below: where e is the number of environments, r is the number of replicates per environment, and 2 G , 2 GE , 2 represent the variance of genotype, interaction between the genotype and environment, and the residual error, respectively (Holland et al. 2003).

Construction of genetic maps and QTL analysis
The ZM175/XY60 population was earlier genotyped with the wheat 55 K SNP array at CapitalBio Technology, Beijing, China (http:// www.capit albio tech.com/).The genetic map constructed with 2437 bin markers from 16,008 SNPs spanned 3,250.7 cM and covered all 21 wheat chromosomes (Luo et al. 2020).The ZM175/LX987 population was genotyped with the wheat 50 K SNP array by the same company.The 50 K array originally contained 66,835 SNP markers but was reduced to 55,224 markers after removal of duplicates.SNPs lacking polymorphism, missing rates > 20%, minor allele frequencies < 25% and redundancy were removed.Linkage analysis and chromosome allocation of linkage groups were completed by the MAP function in IciMapping v4.1 (Li et al. 2007).The final genetic map was constructed using 1217 bin markers from 11,377 polymorphic SNPs spanning 3094.5 cM across all 21 wheat chromosomes.QTL analysis was conducted using the inclusive composite interval mapping (ICIM) method in IciMapping v4.1 (Li et al. 2007).The logarithm of odds (LOD) score was set at 2.5 to claim significance.QTL in different environments with overlapping intervals were considered the same, and QTL detected in at least three environments for ZM175/XY60 and two environments for ZM175/LX987 were regarded as stable.QTL in the suppressed recombination region of chromosome 2DL were analyzed using a general linear model (GLM) in TASSEL v5.0 (Bradbury et al. 2007; http:// www.maize genet ics.net/ tassel).Local Manhattan plots were generated using the CMplot package in R v3.5.1 (https:// github.com/ YinLi Lin/R-CMplot).Linkage disequilibrium (LD) heatmaps were constructed using the LDheatmap package in R v3.5.1 (https:// sfust atgen.github.io/ LDhea tmap).Genetic maps were drawn using MapChart v2.3 (https:// www.wur.nl/ en/ show/ Mapch art.htm).The physical positions of stable QTL were determined by closely linked markers based on RefSeq v1.1 (IWGSC 2018; https:// urgi.versa illes.inra.fr/ blast_ iwgsc/).

Design and validation of KASP markers
SNPs from chip-based markers and candidate gene sequences were converted to chromosome-specific KASP markers for QTL validation.KASP primers were designed using PolyMarker (Ramírez-González et al. 2015; http:// www.polym arker.info/).The FAM (5′ GAA GGT GAC CAA GTT CAT GCT 3′) and HEX (5′ GAA GGT CGG AGT CAA CGG ATT 3′) splice sequences were added to the 5' ends of the alternative competitive primers.PCR was performed using the protocol described in Xu et al. (2019).Consistency between KASP-and original chip-based genotypic results was investigated, and t-tests were conducted to validate the effectiveness of the QTL using KASP markers.

QTL validation
SNP marker AX-109500250 (577.3Mb) tightly linked to the 2DL QTL (562.7-596.2Mb) in the ZM175/XY60 population was converted to KASP marker K_AX-109500250 (Table S1, Fig. 2).The t-tests in the ZM175/LX987 population showed that the mean values of GZnC, GFeC and TGW of lines carrying the ZM175 allele were significantly lower than those for lines carrying the LX987 (XY60) allele (Fig. 2).However, contrasting genotypes of the K_ AX-109500250 marker in the cultivar/line panel were not associated with mean values of GZnC, GFeC and TGW in all environments, indicating that the marker was too far away from the causative gene (Table S7, Table S8).
SNP marker AX-110119937 (460.8Mb) closely linked to the 6DL QTL (458.8-464.6Mb) for GZnC in both the ZM175/XY60 and JD8/AK58 populations was converted to KASP marker K_AX-110119937 (Table S1, Fig. 3).A few accessions (12.1%) of the cultivar/line panel carried the favorable allele (Table S7).The mean GZnC of accessions with the favorable allele were 3.9-6.6%,significantly higher than those with the unfavorable allele in three environments and BLUE, indicating that this marker was close to the causative gene (Table S8, Fig. 3).

Candidate genes and gene-based markers for stable QTL
Suppressed recombination in the chromosome 2DL QTL region was observed in both the ZM175/XY60 and ZM75/ LX987 populations.Since both the ZM175/XY60 and ZM175/LX987 populations had no polymorphism in 596.2-613.1 Mb and 549.7-570.5 Mb regions, respectively (Fig. 4), the 2DL locus was assigned to 570.5-596.2Mb.
Based on gene annotations from RefSeq v1.1 and our earlier consensus genetic map underlying Zn and Fe homeostasis on chromosome 2DL (Tong et al. 2020), TaIPK1-2D (593.5 Mb) was the best candidate gene for the QTL.Sequencing of TaIPK1-2D indicated that XY60 and LX987 shared the same gene sequence, whereas ZM175 was different (Fig. S3).This was consistent with the QTL results that both XY60 and LX987 contributed a favorable allele for GZnC and GFeC.The cultivar/line panel was genotyped using KASP marker K_TaIPK1-2D derived from TaIPK1-2D.The t-tests showed significant differences between GZnC, GFeC and TGW values for these genotypes in at least three environments for each trait (Table S7, Table S8), supporting the prediction of TaIPK1-2D as the candidate gene.TaGS2-2D (595.2Mb) associated with grain weight in the 2DL QTL region was also sequenced; XY60 and LX987 shared the same sequence that differed from the sequence in ZM175, indicating that TaGS2-2D was a promising candidate gene for the TGW locus (Fig. S4).The cultivar/line panel was genotyped using KASP marker K_TaGS2-2D derived from TaGS2-2D, and t-tests showed significant differences for GZnC, GFeC and TGW between the two opposite contrasting alleles in some environments (Table S7, Table S8).Since recombination in the region was suppressed, TaIPK1-2D for GZnC/GFeC and TaGS2-2D for TGW physically about 1.7 Mb apart were co-segregating in both RIL populations (Fig. 4).
The 4BS QTL were co-located with RHT-B1 when its KASP marker was added to the chromosome 4B genetic map (Fig. 5a).The t-tests showed that Rht-B1a significantly increased GZnC and GFeC supporting its candidate role for the 4BS QTL (Table S9).The 6A QTL was also associated with plant height and it was located at a suppressed recombination interval of 73.4-445.4Mb spanning the centromere (Fig. 5b).Four reduced height genes (Rht14, Rht18, Rht24 and Rht25) have been reported on chromosome 6A (Ford et al. 2018;Haque et al. 2011;Tian et al. 2022;Zhang et al. 2023).Among them, Rht14 and Rht18 were obtained from artificial mutagenesis, whereas Rht24 and Rht25 were in natural resources.Sequencing results showed that ZM175, XY60 and LX987 all contained the Rht24b allele (Fig. S5).Haplotype analysis of Rht25 in the parents showed that ZM175 and LX987 carried Rht25d the semi-dwarf allele of the gene, while XY60 carried the Rht25a the wild-type allele (Fig. S6), and Rht25a was therefore likely the candidate gene for the 6A QTL.
The 6DL QTL was located in the interval 458.8-464.6Mb (Fig. 3).According to gene annotations for RefSeqv1.1 and typical genes associated with Zn and Fe homeostasis, TraesCS6D02G382900 (462.4Mb), a nicotianamine synthase (NAS) gene is located in this interval (Fig. 3).Based on the previously named TaNAS1 to TaNAS9 in wheat (Bonneau et al. 2016) we named the candidate gene TaNAS10-6D.The TaNAS10-6D sequences in XY60 and JD8, with the favorable allele(s) for GZnC were identical whereas the sequences in ZM175, and AK58 the contrasting unfavorable allele carriers, together with LX987, were also identical.These data and QTL results supported the prediction that TaNAS10-6D was a likely candidate gene for the 6DL QTL (Fig. S7).When the gene-based KASP marker K_TaNAS10-6D was used to genotype the cultivar/line panel and t-tests showed a significant difference between contrasting two alleles for GZnC in at least two environments (Table S7, Table S8).The gene expression profile showed that TaNAS10-6D was specifically expressed in roots (Fig. S8), consistent with the expression profile of class I nicotianamine synthase genes (Bonneau et al. 2016), and further supported its candidacy.

Discussion
Grain Zn and Fe concentrations are complex quantitative traits and are greatly impacted by environmental factors.Soil homogeneity and phenotypic evaluation in multiple environments are key to obtaining reliable results.In this study, there were four co-located GZnC/GFeC loci, and four independent GZnC and two independent GFeC loci stably detected in two RIL populations with ZM175 as a common parent.In addition, two gene-based KASP markers were successfully developed for the 2DL QTL for GZnC/GFeC and the 6DL QTL for GZnC.

Genetic variation, heritability and correlation coefficients of GZnC, GFeC and TGW
The ZM175/XY60 and ZM175/LX987 RIL populations showed a large variation for GZnC, GFeC and TGW with near normal distributions (Table 1, Fig. S1), indicating that these traits are quantitatively inherited.This study is consistent with previous reports showing the polygenic nature of GZnC, GFeC and TGW (Crespo-Herrera et al. 2016;Hao et al. 2014).In our study, the heritability values for GZnC, GFeC and TGW were high (Table 1), indicating that these traits are affected by genetic factors.Heidari et al. (2016) reported high heritability values for GZnC and GFeC, and Heidari et al. (2011) and Tahmasebi et al. (2016) reported high heritability values for TGW.Similarly, Hao et al. (2014) reported high heritability values of 0.79 and 0.78 for GZnC and TGW, respectively.GZnC and GFeC were positively correlated with TGW, and GZnC and GFeC were positively correlated with each other (Table S4), suggesting that these positively correlated traits may be helpful for improving more than one trait simultaneously.However, some studies reported that GZnC and GFeC were not correlated with TGW (Crespo-Herrera et al. 2016;Heidari et al. 2016;Liu et al. 2019), indicating the complexity of trait relationships in some specific genetic backgrounds.

Independence to dilution effect
The introduction of "Green Revolution" genes, in particular Rht-B1b causing reduced height, greatly improved dry matter accumulation in favorable environments, but the redistribution of minerals from the vegetative tissue to grain did not equate with the redistribution of photosynthates, causing a "dilution effect" (Fan et al. 2008;Murphy et al. 2008).The common parent ZM175 carried the wild-type allele (Rht-B1a) at the RHT-B1 locus, whereas XY60 and LX987 carried Rht-B1b.RILs with the semi-dwarf allele usually had more spikes per unit area and were therefore more productive than the RILs with the wild-type allele but had lower GZnC, GFeC and TGW (Luo et al. 2020;Velu et al. 2017).Crespo-Herrera et al. (2016) also reported a pleiotropic QTL for GZnC, GFeC and TGW on chromosome 4BS that was positively correlated with plant height.Similar results were found at a chromosome 6A locus (Wang et al. 2021;Xu et al. 2014).In addition, meta-QTL analysis showed that QTL for GFeC and yield-related traits were co-located at RHT-D1 locus (Shariatipour et al. 2021b).
A QTL on chromosome 2DL showed pleiotropic effects on GZnC, GFeC and TGW that were independent of plant height (Fig. 2).This QTL was likely a new GZnC/GFeC locus (Tong et al. 2020).Previous studies reported QTL for TGW (Cui et al. 2014(Cui et al. , 2016;;Liu et al. 2014a, b;Su et al. 2018), spike number per plant (Cui et al. 2014), and grain number per spike (Cui et al. 2014;Shi et al. 2017) at the same 2DL locus using different populations.Moreover, for the same ZM175/XY60 RIL population, agronomic traits including spikelet number per spike, yield per plant, aboveground biomass per plant, harvest index, and grain perimeter, were all associated with a common QTL (Luo et al. 2020).TaGS2-2D (595.2Mb) was predicted as candidate gene affecting these agronomic traits (Hu et al. 2018;Li et al. 2011).The nearby gene TaIPK1-2D (593.5 Mb) involved in biosynthesis of phytic acid, a major antinutrient for mineral absorption (Aggarwal et al. 2018), was a likely gene candidate for GZnC and GFeC.The pleiotropic effect of 2DL locus was probably caused by suppressed recombination in the gene region.Chromosomal segments with suppressed recombination are commonly found in modern cultivars and founder parents; and superior haplotypes of these regions tend to be selected during breeding (Hao et al. 2020).The 2DL QTL carrying favorable alleles of TaIPK1-2D, TaGS2-2D and probably other important genes in XY60 and LX987 represent a useful chromosomal segment for wheat breeding for both yield and nutrition.Hao et al. (2014) also detected a pleiotropic QTL on chromosome 2B that was associated with GZnC and TGW and independent of plant height.Apart from these pleiotropic QTL, QTL on chromosomes 1BL, 1D, 5DL, 6DL and 7BL were independent of plant height and TGW and might not cause dilution effects when selected.

QTL detection in multiple related populations
Two RIL populations with ZM175 as a common parent were used to detect QTL for GZnC and GFeC in this study.Among four QTL with favorable alleles from ZM175, only the 4BS QTL was detected in both populations; detection of the other three (QGZnzx.caas-5DL,QGFezl.caas-5DL and QGZn/Fezl.caas-7BL)was restricted to a single population.The tightly linked markers for QGZnzx.caas-5DLwere not polymorphic between ZM175 and LX987, and QGFezl.caas-5DL was not polymorphic between ZM175 and XY60, possibly explaining the difference in QTL detection between the two populations.Although the markers flanking QGZn/Fezl.caas-7BLwere polymorphic between ZM175 and XY60, that QTL was not detected in the ZM175/XY60 population probably because the markers were not close enough or its effect was masked by other QTL.The current results show that for complicated traits such as GZnC and GFeC, it is better to use related mapping populations for QTL analysis to achieve a clearer representation of the genetic architecture.

Application in wheat biofortification
Since GZnC and GFeC in modern leading wheat cultivars are generally lower than in pre-"Green Revolution" genotypes (Velu et al. 2019), it is necessary to search for more genetic variation and to develop effective molecular markers for use in breeding for higher GZnC and GFeC in wheat.A good marker for marker-assisted selection (MAS) should be one that is linked with QTL identified in multiple locations, years, and environmental conditions.This ensures the marker's robustness and reliability across different genetic backgrounds and growing environments.In the present study, we analyzed the genetic basis underlying higher GZnC and GFeC in wheat cultivar XY60 compared to the modern cultivars ZM175 and LX987 and converted SNPs linked to QTL detected in multiple environments on chromosomes 2DL and 6DL to KASP markers.The KASP marker on chromosome 2DL (K_AX-109500250, 577.3 Mb) was validated in the ZM175/LX987 population, but not in a panel of cultivars/lines.which might be related to the high genetic diversity and abundant recombination events in the cultivar population, and K_AX-109500250 was far away from the causative gene.The KASP marker on chromosome 6DL (K_AX-110119937, 460.8 Mb) might be very close to the causative gene that was verified in both a JD8/AK58 RIL population and the cultivar/line panel.A small portion of accessions (12.1%) in the panel carried the favorable 6DL allele, indicating that it has not been widely used in breeding.Previously, we identified several wheat orthologs of known Zn and Fe homeostasis genes in model plants through genome-wide association studies, but the lack of functional markers hindered breeding for biofortification (Tong et al. 2020(Tong et al. , 2022)).In this study, two gene-based KASP markers K_TaIPK1-2D and K_TaNAS10-6D were validated in the panel of 343 wheat cultivars/lines supporting their potential usefulness in marker-assisted selection.A genotyping by target sequencing (GBTS) platform showed potential for developing markers and rapid identification of functional genes (Guo et al. 2021).Thus, a candidate gene-based breeding strategy using GBTS for GZnC and GFeC will pave a new path to pyramid favorable genes for wheat biofortification.
217 Page 14 of 16 mapping; JT, YW, ZP and JZ participated in phenotyping; JT, YW, JS, YZ, LL and AZ participated in field trials; and YH, AR, ML, SC, XX and ZH reviewed and revised the manuscript.

Fig. 2
Fig. 2 Genetic map of the major QTL on chromosome 2DL in the ZM175/XY60 RIL population and its validation in the ZM175/ LX987 RIL population.a QTL on 2DL based on the best linear unbiased estimation (BLUE).Effect of KASP marker K_AX-109500250

Fig. 3
Fig. 3 Genetic map of QTL on chromosome 6DL for grain Zn concentration in the ZM175/XY60 and JD8/AK58 RIL populations and its validation in a cultivar/line panel.a QTL identified on 6DL in the ZM175/XY60 RIL population.b QTL identified on 6DL in the JD8/ AK58 RIL population.c Physical positions of flanking markers and

Fig. 4
Fig. 4 The suppressed recombination region on chromosome 2DL.a Local Manhattan plot and linkage disequilibrium heatmap of single nucleotide polymorphisms within the 2DL QTL in the ZM175/XY60 RIL population.b Local Manhattan plot and linkage disequilibrium heatmap of single nucleotide polymorphisms within the 2DL QTL in the ZM175/LX987 RIL population.The flanking markers of the sup-

Table 2
Stable QTL for GZnC, GFeC and TGW identified in the ZM175/XY60 RIL population Trait a QTL Environment b Position (cM) Left marker Right marker Physical interval (Mb) LOD c PVE (%) d Add e

Table 2
(continued) d LOD-logarithm of odds e Add-the additive effect, positive or negative values indicate that the alleles were inherited from ZM175 or XY60 Theoretical and Applied Genetics (2023) 136:217 1 3 217 Page 8 of 16

Table 3
Stable QTL for GZnC, GFeC and TGW identified in the ZM175/LX987 RIL population Add-the additive effect, positive or negative values indicate that the alleles were inherited from ZM175 or LX987 a GZnC-grain Zn concentration; GFeC-grain Fe concentration; TGW-thousand grain weight b GY-Gaoyi; GC-Gaocheng; BJ-Beijing; XJ-Urumqi; BLUE-best linear unbiased estimation c PVE-phenotypic variation explained d LOD-logarithm of odds e