QTL underlying iron toxicity tolerance at seedling stage in backcross recombinant inbred lines (BRILs) population of rice using high density genetic map

Fe is a trace element considered to be essential for rice, and it drives several metabolic processes. Fe toxicity occurs due to excessive Fe ions (Fe) and which, disturb cellular homeostasis and dramatically reduces the rice yield. A set of 118 BRILs made from a cross of japonica cv.’02428’ and indica ‘Changhui 891’ was used with high density bin map constructed by using high quality SNP to identify the QTL for Fe toxicity tolerance. As a whole total of 23 QTL were identified for various seedling traits, 3 under control with phenotypic difference ranging from 14.21% to 62.46%, 11 QTL under stress with phenotypic difference ranging from 7.89% to 47.39% and 9 under stressed/control ratio with phenotypic variance ranging from 9.17% to 183.50%. LOD values of QTL ranging from 4.05 to 17.04 in control, 3.41 to 8.09 in stress and 2.84 to131.63 in stress/control ratio. Shoot length (SL), root length (RL), shoot fresh weight (SFW), root fresh weight (RFW), shoot dry weight (SDW), and root dry weight (RDW), were used to estimate the degree of Fe tolerance. Many stable QTL, qSSDW-4, qSSDW-6, qRSDW-4 and qRSDW-6 affecting SDW were detected and beside this some new QTL, qRSFW-1, qRRFW-10 and qRRDW-1 were successfully identified significantly contributing to Fe toxicity tolerance in rice. The results of current study indicated that these novel regions could be transferred via markers assisted section and QTL pyramiding to develop Fe resistant lines in rice.


3
Generally, rice is categorized in two groups (indica and japonica) based on large genetic variability. Many researchers have studied large phenotypic diversity among japonica and indica rice varieties (Liu et al., 2005), as japonica is more resistant to Fe toxicity than indica. Use of resistant genotypes is one of the best strategy to improve rice yield on soils effected by Fe toxicity (Audebert, 2006). Fe toxicity tolerance is appear to be polygenic inherited traits and therefore QTL mapping along with marker assisted selection is most reliable technique to improve genotypes resistance (Mackill et al., 1999). Earlier screening studies demonstrated that , environmental conditions, Fe stress level, screening systems played a key role to determine the response of rice cultivars to Fe stress (Wu et al., 2014).Therefore it is important to breed the special varieties adopted to specific environment and those adopted to wide range of environments. Several genes controlling Fe uptake, transport and accumulation belong to five protein groups have been identified, OsNRAMPs, OsFROs, OsZIPs, OsFERs and OsYSLs (Chandel et al., 2010). Identification of specific traits involved in contributing Fe toxicity tolerance would be the most promising step for reliable mapping studies and successful identification of Fe tolerant QTL in the environments diverse in Fe toxicity patterns. The cultivars 'EPAGRI 108', 'BR-IRGA 409' and japonica cultivar '02428' are Fe tolerant cultivars ( Stein et al., 2019;Rasheed et al., 2020c).
Various QTL against Fe toxicity tolerance have been identified in rice using population derived from two parents (Dong et al., 2006;Dufey et al., 2009) using easily measurable parameters, such as SL, RL, SDW and RDW, because it is difficult to measure physiological traits at whole population level (Dufey et al., 2015b).
Numerous mapping and screening experiments have been conducted to study Fe toxicity tolerance with different tolerant response in rice. A japonica variety, 'Azucena' was studied for its tolerance by using (250 mg L −1 Fe 2+) for duration of 4 weeks using hydroponics conditions and marked as tolerant (Dufey et al., 2009) but it showed susceptibility when exposed to (1,500 mg L −1 Fe 2+ ) (Engel et al., 2012). Contradictory performance of same genotype under different Fe toxic levels provided an opportunity to understand the genetics basis of iron toxicity tolerance. Lot of studied have been conducted in past to identify the QTL behind Fe toxicity tolerance (Dufey et al., 2009;Dufey et al., 2012;Fukuda et al., 2012). Due to strong interaction of genotype and environment in field, the testing of genotype tolerance against iron toxicity in hydroponic condition is more rapid and reliable way. Few studies were conducted using genome wise association analysis for complex traits related to iron tolerance (Matthus et al., 2015). Comprehensive understanding of genetic mechanism of Fe toxicity tolerance is important to speed up the development of Fe tolerant varieties.
Many QTL for rice tolerance to Fe toxicity have been identified by earlier researchers. A total of 29 QTL in rice for Fe toxicity tolerance using GWAS technique have been identified for SL, RL, SFW, RFW, SDW and RDW (Zhang et al., 2017). Dufey et al. (2015a) conducted QTL analysis using 220 BC3DH lines tested in hydroponics condition in presence or absence of Fe toxic levels and reported 28 QTL for various morphological traits. Similarly Liu et al. (2016) used two reciprocal introgression lines with same parents to identify QTL against Fe toxicity tolerance. They have successfully identified 9 QTL for Fe toxicity tolerance using several morphological indicators such as, RDW, SDW, and total dry weight (TDW). Fe toxicity is still a challenge for rice breeders to sustain its production on Fe toxic soils by identifying putative QTL, because until now this mechanism is not fully understood (Nughara et al., 2016;Zhang et al., 2017). Exploitation of appropriate mapping population is a reliable and effective strategy in QTL analysis, because of the fact that standard segregating populations cannot give accurate knowledge about size and position of individual QTL, mainly QTL with small effect, due of genetic background noise (Jiang et al., 2017). In such circumstances, use of near isogenic lines (NIL) and chromosomal segment substitution lines (CSSL) can be ideal to overcome this problem and therefore considered as ideal for identification of QTL ( Kubo, 1999;Bian et al., 2010;Jiang et al., 2017), but using these population are time consuming and increase labor cost, which limited the rapidity of gene cloning. BRILs population holds higher percentage of genome from recurrent parent and easily crossed with recurrent parent to develop NIL population for achieving the cloning of targeted QTL. Hence BRILs is considered a permeant and stable population and easier to develop than that of CSSL and NILs population. With the passage of time idea of developing BRILs population gained more attention for genetic analysis of polygenic traits in rice against various abiotic stresses ( Wang et al., 2011;Hosseini et al., 2012). 4 QTL mapping is one of the most reliable and powerful technique to locate gene of interest on chromosome (Nughara et al., 2016) and mapping of QTL on chromosome depend on marker density in linkage group as well as size of QTL interval (Da et al., 2000). Idea of developing reliable markers gained more and more attention with passage of time by researchers to construct high density linkage map. Use of restriction fragment length polymorphism (RFLP) and single sequence repeat (SSR) markers do not meet our needs of QTL identification when linkage distance is zero, therefore use of additional markers would be better to enhance QTL mapping resolution (Liu et al., 2008). Use of SNP would facilitate the probability of increase in high resolution QTL mapping. Use of SNP is more beneficial as comparted to other markers because old markers like, SSR, RFLP are not appropriate when their linkage distances are zero and cannot meet the necessities of the high QTL mapping resolution . Use of SNP markers by next generation sequencing tools will rise the possibility to rise resolution of QTL mapping (Jiang et al., 2017). In this study high resolution bin-map was constructed using re-sequence strategy for recognition of QTL for Fe 2+ tolerance. Aim of currently study was to evaluate a set of 118 BRIL lines population under hydroponic conditions to determine the degree of Fe toxicity tolerance in BRILs and to recognize the putative QTL controlling Fe tolerance.

Population development
A BRILs population consists of 118 lines and its parents were used to estimate the degree of Fe toxicity tolerance at seedling traits and to identify the putative QTL linked with Fe toxicity tolerance. Population was made from a cross of '02428' (japonica) and 'Chunghui/891' (indica) parents at experimental station of Jiangxi Agricultural University and Linwang of Hainan Province. Parents were crossed and population was allowed to self-fertilization for six generations to produce BC1F6.
Screening for Fe tolerance using hydroponic experiment Experiment was conducted at experimental station of Jiangxi Agricultural University, Nanchang PR. China during 2019. A total of 118 rice BRILs along with their parents were evaluated in culture solution experiment to investigate the consequence of Fe toxicity on each line and parent seedlings. Experiment was laid out as RCBD with three replications per block, 10 seedlings from (3-4) lines per replication. Firstly seeds of the parents and BRILs population were disinfected with 1% H2O2 for 30 minutes and rinsed several times with distilled water (Satoh et al., 2016) and soaked in de-ionized water for oven night than grown on filter papers placed in petri plates. One BRILs line was grown in three dissimilar plates alongside with its parents; hence float in 0.5 mmol l −1 CaCl2 (pH 5.0) solution. Simple CaCl2 solution could be used to screen at young seedling stage when seed can provide vital nutrients and to escape precipitation in Yoshida solution.
The whole experiment was conducted in culture room with temperature 27±2 °C, 16 hours of light and relative humidity of 70%. Seeds were allowed to germinate at this temperature for 4 days, and on 4 th day 10 seedlings with identical size were selected and transferred in plastic trays protected PVC sheet containing nylon screen attach holes. On 5 th day CaCl2 at concentration of 0.5 mmol −1 (as control) and Fe 2+ in form of FeCl2 at concentration of 100 μM (as stress) were applied to the seedlings in nutrients solution following method of (Yoshida et al., 1976). We have conducted a preliminary experiment and treated rice lines with different doses of Fe toxicity and we have concluded that 100 μM is more effective dose as plant showed large variation towards Fe toxicity on this concentration. The pH of the solution during the experiment was adjusted to 5.0 using I N NAOH/HCL (Zhang et al., 2017). The nutrients solution of the experiment was replaced after one weak and seedlings were exposed to stress for a period of 2 weeks to determine the effect of Fe toxicity on seedling traits. We have used the CaCl2 to prepare the solution for control and we used the FeCl2 to prepare the Fe stress solution. FeCl2 was source of Fe stress and was applied at concentration of100 μM per liter.

Phenotypic data collection
After 15th day of treatment, when significant differences among the control and stress treatments were clearly noticed than, 10 plant samples from each parent and line were harvested to determine the SL and RL. SL was measured from the longest leaf to the base of shoot with the help of ruler and longest root was measured from longest root to the base of root (Nughara et al., 2016). After this root and shoot were detached to measure the SFW and RFW with the help of an electronic balance. The samples of shoot and root were than oven dried at 60 °C for 72 hours to determine the SDW and RDW. The relative values like, relative shoot length (RSL), relative root length (RRL), relative shoot fresh weight (RSFW), relative root fresh weight (RRFW), relative shoot dry weight (RSDW) and relative root dry weight (RRDW) were taken by using formula (trait value in treatment/trait value in control) x 100 (Nughara et al., 2016).

Data analysis
The phenotypic data was subjected to the analysis of variance using Minitab v.18 software (Pennsylavania state University, PA, USA) based on general linear model at p < 0.05 level and Tukey honest significant difference (HSD) test was used for comparison of the mean values. SPSS software version 20 (SPSS, Chicago, IL, USA) was used to carry out further analysis, descriptive statistics, Pearson correlation, frequency distribution, histogram and box plot. The broad sense heritability was calculated using the formula: h 2 =Vg (Vg+Ve/r), where, (Vg) genotypic variance, (Ve) error variance and r is replication.

QTL analysis in BRILs population
A high-density linkage map was developed by using SNP with high quality (Jiang et al., 2017). SNP used in this study are more stable and consistent markers. QTL analysis was carried out using the software IciMapping v 4.1 software (http://www.isbreeding.net). Significance of the QTL for evaluated traits was claimed using LOD value of >2.50 (Bian et al., 2015;Jiang et al., 2017). Technique of McCouch (McCouch, 2008) was applied for QTL nomenclature.

Phenotypic variation and traits correlation
Analysis of variance was performed for all investigated traits of 118 BRILs across different experimental conditions. Different range of variation of all investigated traits was observed across both control and stress conditions (Table 1). Higher estimates of heritability were recorded for all traits, indicating the large scope of selection present for all traits. Regarding heritability estimates all traits showed varying pattern, SL (93.81%), RL (96.88%), SFW, (96.31%), RFW, (93.44%), SDW (92.97%) and RDW (69.06%) ( Table 3). Higher estimates of broad sense heritability revealed that all of these traits are highly heritable and selection of genotypes based on these traits would be highly appreciable. Mean values of SL, SFW and SDW under control conditions were considerably decreased as compared to stress indicated that these are least affected by the Fe toxicity under hydroponic conditions. Mean values of RL and RFW under stress conditions were considerably reduced compared to control (Table 2)  SL, shoot length; RL, root length; SFW, shoot fresh weight; RFW, root fresh weight; SDW, shoot dry weight; RDW, root dry weight; SOV, source of variation; DF, degree of freedom; SS, sum square; MS, mean square; H2, heritability; * shows significant at P ≤ 0.05 and ** shows highly significant at P ≤ 0.01.
Under control condition positive non-significant correlation was observed between RL and SL (0.19), while SFW also exhibited non-significant positive correlation with SL (0.31) and RL (0.15) ( Table 3). RFW showed negative non-significant correlation with SL (-0.01), positive non-significant correlation with RL (0.14) and highly significant positive correlation with SFW (0.70**). SDW exhibited non-significant positive correlation with SL (0.32), RL (0.35), RFW (0.35) and highly significant positive correlation with SFW (0.68**). RDW demonstrated non-significant positive correlation with SL (0.01), RL (0.35), SDW (0.43), significant positive correlation with SFW (0.50*), and highly significant positive correlation with RFW (0.78**) and under control. Significant correlation demonstrated that these traits could be improved via direct selection. Under FeCl2 RL exhibited non-significant positive correlation with SL (0.33) and, SFW showed significant positive correlation with SL (0.48*) and positive non-significant correlation with RL (0.30). RFW had negative non-significant correlation with SL (-0.05), positive non-significant correlation with RL (0.11)  (Table 3), all the variables which showed significant and highly significant positive correlation could be selected to improve direct selection.  Identification of the QTL for seedlings traits of BRILs in response to control and Fe stress A total of 14 QTL in control and Fe stress conditions were identified for seedling traits in BRILs population (Tables 4 and 5, Figures 1 and 2), including 3 QTL detected only under control environment and 11 detected only in Fe stress condition. Three QTL underlying three traits, CRL, CRFW and CRDW, specifically one QTL, qCRL-1 for CRL, one QTL, qCRFW-3 governing CRFW and one QTL, qCRDW-3 underlying CRDW were identified under control experiment on chromosome 1 and 3 respectively. The QTL, qCRDW-3 was identified in both control and stress environment, which revealed its background independence. Range of LOD values and phenotypic variances shown by these QTL were 4.05, 3.31, and 17.04 and 14.36%, 14.21% and 62.46% respectively. For three traits, CSL, CSFW and CSDW no QTL were mapped on any chromosome in control background. No QTL was detected on chromosome, 2, 4, 5, 6, 7, 8, 9, 10, 11 and 12 under control experiment. No QTL was detected for CSL, CSFW, and CSDW in control condition.

Traits
Treatments Under Fe stress 11 QTL affecting four seedling traits, SSFW, SRFW, SSDW, and SRDW were detected on chromosomes 1, 3, 4, 6, 10. Two QTL, qSSFW-6, qSRFW-10 governing SSFW and SRFW were identified on chromosomes 6 and 10 among the flanking markers (bin6-159, bin6-160 and bin10-161, bin10-162), with LOD values 3.41 and 3.94 and accounting for 12.26% and 13.52% of phenotypic variances respectively. Allele for qSSFW-6 was shared by male parent as shown by positive additive effect and increased trait value. Two QTL, qSSDW-4 and qSSDW-6, affecting SSDW were detected in Fe stress experiment on chromosome 4 and 6, with LOD value of 6.66 and 2.64, accounting 21.42% and 7.89% of phenotypic variance, respectively. QTL, qSSDW-6 had positive additive effect, showed that allele at this locus was shared by donor parent and increases index value. Negative additive effect showed that female parent is involved in increasing SSDW of BRILs population under stress environment. Seven QTL for SRDW were identified in Fe stress on chromosome 1 and 3, with LOD values ranging from 6.79 to11.46 and phenotypic variances 47.12% to 47.39%, respectively.

Discussion
Use of BRILs population enables the genetic analysis of QTL Use of suitable mapping population is a reliable and effective strategy in QTL analysis, because of the fact that standard segregating populations cannot give accurate knowledge about size and position of individual QTL, mainly QTL with small effect, due of genetic background noise (Jiang et al., 2017). In such circumstances, use of NILs and CSSL can be ideal to overcome this problem and therefore considered as ideal for identification of QTL ( Kubo, 1999;Bian et al., 2010;Jiang et al., 2017), but using this populations is time consuming and increase labor cost, which limited the rapidity of gene cloning. BRILs population holds higher percentage of genome from recurrent parent and easily crossed with recurrent parent to develop NIL population for achieving the cloning of targeted QTL. Hence BRILs is considered a permeant and stable population and easier to develop than that of CSSL and NIL population. With the passage of time idea of developing BRILs population gained more attention for genetic analysis of polygenic traits in rice against various abiotic stresses (Hosseini et al., 2012;Wang et al., 2011). Currently a population of 118 BRILs was developed by crossing japonica ('02428')/ 'Changhui/891', hence; we propose that, use of this population would be a better choice to understand the complicacy of important quantitative traits in rice.

Relationship of currently identified and previously reported QTL
For successful identification of QTL an effective nutrient solution with toxic levels of Fe is required (Liu et al., 2016). Fe stress tolerance in rice can be characterized at morphological, biochemical and physiological levels (Blum, 2011). Fe tolerance in rice is mainly contributed by morphological attributes and these traits are polygenic in nature (Tiwari and Mamrutha, 2013). Crop tolerance to any stress is influenced by the environment to a great extent, hence previously it was proposed to improve crop tolerance to Fe stress for morphological traits using MAS selection (Zhang et al., 2017). All of the secondary traits are linked with stress 13 tolerance and should have following criteria (1) genetically associated with Fe stress tolerance in Fe stress condition; (2) their heritability should be high in screening system; (3) lines should possess large diversity for given traits (4) Traits should be measured rapidly and they should be economic (Lafitte and Courtois, 2000). Currently we have measured Fe stress tolerance related to morphological attributes i.e. SL, RL, SFW, RFW, SDW and RDW under control, stress and stress/control ratio. QTL detected for the ratio traits measured in stress/control experiment were the major QTL that directly contributed to Fe toxicity tolerance in rice (Yue et al., 2006), but traits variation can be reduced as a result some of the QTL may be un-identified (Zhang et al., 2017). Hence, comparison of QTL detected under control and stress should also be done. All of the detected QTL in our study were associated with seedling traits, RL, SL, SFW, RFW, SDW and RDW. Of the 23 QTL identified, 3 were detected under control, 11 under stress and 9 under stress/ratio.
Only one QTL, qRSL-12 affecting RSL was identified in current study under stress/control condition accounting for 9.92% of phenotypic variance. Meng et al. (2017) used five MAGIC populations against Fe toxicity and identified four QTL influencing RSL out of them one QTL qRSL-12 with 8.1% variance was common to our reported QTL contributing to Fe toxicity tolerance in rice. However, Zhang et al. (2017) previously presented contradictory results and identified two QTL, qCSL-12, qSSL-12 affecting CSL and SSL on chromosome 12 in both control and Fe stress condition with 4.4% and 6.9% of phenotypic variance.
Similarly, Zhang et al. (2013) used reciprocal advanced backcross introgression lines and identified 42 QTL, out of them 17 QTL were reported for SL on chromosomes 1,2,3,5,6,7,8,9,11 not common to our detected QTL. This difference is mainly due to use of different populations and markers. RL is important seedling trait that is easily to measure for Fe toxicity tolerance. Currently, only one QTL, qCRL-1 is mapped on chromosome 1 under control experiment. Three QTL, qRL-6, qRL-7 and qRL-8 were identified by Meng et al. (2017) for RL differed from our reported QTL due to use of different populations and nutrient solution. Zhang et al. (2017) detected, qCRL-2 for CRL in control, and qSRL-3 for SRL in Fe stress condition, which were in contrast with our identified QTL. By comparing with previous results, currently identified QTL, is not reported before.
So, what QTL RSFW-1 detected in current experiment is newly reported QTL with the allele from recurrent parent. This region is strongly affects the stress tolerance index of SFW. Zhang et al. (2013) detected 42 QTLs for Fe and zinc toxicity tolerance across all environments but, they have not reported any QTL for SFW. Regarding RFW we have detected three QTLs, qCRFW-3, qSRFW-10, qRRFW-10, one under control, one under stress and one under stressed/control ratio experiment with 14.21%, 13.52% 9.17% of phenotypic variances. Previously no QTL was reported for RFW in any study. However, RFW is an important parameter in studying Fe toxicity tolerance. These novel regions may contribute to Fe toxicity tolerance in BRILs and could be cloned and transfer via MAS strategy.
Fe toxicity strongly influences the SDW in hydroponic condition, treated with toxic level of Fe. Most of the QTL, reported in previous studies were mainly related to SDW and RDW. It is of great interest that four QTL, two qSSDW-4, qSSDW-6 in Fe stress and two qRSDW-4, qRSDW-6 in stressed/control environment were identified in this study on chromosome 4 and 6 with 21.42%, 7.89%, 19.23%, and 11.94% of variance. These regions were identified on same chromosomes but among different markers. QTL on chromosome 6 in both the environments had positive additive effect which indicated that at this locus the allele was shared by donor parent. In many studies that QTL for SDW have been detected this revealed that this morphological trait is largely contributing to Fe toxicity tolerance. Dufey et al. (2012) identified a QTL, qSDW-4 for Fe toxicity stress on chromosome 4 accounted for 10.9% of phenotypic variance.
This genetic factor is on the same chromosome as reported in current experiment. These finding revealed that chromosome 4 may provide important regions that contributed to Fe toxicity tolerance. Zhang et al. (2017) identified five QTL influencing SDW on chromosome 3, 6, 11, 12. QTL, qSSDW-6 detected under Fe stress and RSDW-6 detected under stressed/control condition accounted for 9.2% and 7.9% of variances were common to the QTL in identified in current study. Zhang et al. (2013) also mapped QTL, qRSDW-6 under Fe stress/control ratio and qCSDW-4 under control condition on chromosome 6 and 4. qRSDW-6 is identical to our detected QTL for RSDW on chromosome 6. Comparison of these regions provided a strong base to isolate and transfer these regions for improving growth under Fe toxic condition.
Similarly, Liu et al. (2016) used two reciprocal IL and identified two QTL for RSDW in Fe stressed/control environment. QTL qRSDW-11-1 and qRSDW-11-2 were mapped on different chromosome 11 in this study. In another experiment conducted by Meng et al. (2017) two QTL, qSDW-2, qSDW-5 were mapped on chromosome 2 and 5 for Fe toxicity tolerance with positive additive effect. These mapped QTL are on different chromosomes and at different positions. In another study conducted by. Dufey et al. (2015a) three QTL for SDW under Fe toxicity condition were mapped on chromosome 3, 5 and 12 with positive additive effect.
Most of the QTL identified in this study are related to RDW, one under control, seven under Fe stress and four under stressed/control ratio. Previously lot of studies has identified QTL related to RDW. QTL, qCRDW-3, qSRDW-3 identified in this study with highest phenotypic variances 62.46%, 47.39% were consistent to previously identified QTL, qCRDW-3, qSRDW-3 reported by Zhang et al. (2017) explained 8.5% and 7.8% variances. However, Wan et al. (2003) identified a QTL, qRDW-3 for RDW at chromosome 3 at the region of C25-C515 with explained variance of 35% and positive additive effect. These results are strongly supported by earlier findings. Five QTL for SRDW were mapped on chromosome 1 under Fe stress and four QTL were mapped for RRDW on chromosome 1 in stressed/control ratio condition in current experiment. In comparison with earlier reports, Dufey et al. (2015a) identified qRDW-1 affecting RDW, Wan et al. (2003) also detected two QTL for RDW on chromosome 1. Results of Liu et al. (2016) are in line with our findings, they identified two QTL, qRRDW-1-1 and qRRDW-1-2 on chromosome 1 in stressed/control ratio, but we identified four QTL on same chromosome due to use of different breeding material and markers. We have identified many new QTL for RDW on chromosome 1 in Fe stress and stressed/control ratio and detection of large QTL on same chromosome is of great interest for Fe tolerance. Hence, RDW is strongly associated with Fe toxicity tolerance in rice. These QTL could be cloned and transfer to develop Fe resistant lines in future studies (Rasheed et al., 2020a(Rasheed et al., , 2020b.
qSSDW-4 and qRSDW-4 were detected on same chromosome 4 across stress and ratio condition, likewise, chromosome 10 had two QTL, qSRFW-10 and qRRFW-10 across Fe stress and stressed/control ratio. Current study identified two stable QTL, qCRDW-3 for CRDW in control and qSRDW-3 for SSDW in Fe stress on chromosome 3 on same position and four stable QTL for SDW across stress and stressed/control ratio. These stable QTL could be isolated and cloned for further implications in plant breeding. Fe toxicity tolerance in difficult to measure because it is complex soil nutritional constraint and Fe toxicity tolerance is complex polygenic trait and effect of toxicity is sometime not clearly visible at seedling traits (Das et al., 2021).
Secondary tolerance index like SL, SFW, SDW, RFW and RDW are easy to measure and can help plant breeder to perform selection for Fe toxicity tolerance. These above-mentioned QTL would be useful to develop the resistant lines against Fe toxicity tolerance.

Conclusions
Fe tolerance is important trait to study in rice. Many putative QTL were identified in current study controlling many seedling traits. QTL, SSDW-4, SSDW-6, RSDW-4, RSDW-6 were stable across both stress and stress/ratio environment. In this study we used novel BRILs population and high-density map for QTL mapping. This is the reason this study is different from other studies. We used many secondary tolerance indices like SL, SFW, RFW, SDW and RDW for estimation of Fe tolerance. Three new QTL, qRSFW-1, qRRFW-10 and qRRDW-1 were identified in our study and these QTL can be used to develop Fe tolerant cultivars using MAS and QTL pyramiding. These findings shed light on comprehensive understanding of Fe tolerance mechanism in rice to sustain rice growth on Fe affected soils.

Authors' Contributions
AR planned and conducted research work, collected data, prepared manuscript. GMW and HK planned research work and data analysis. AMS helped in collection of data. MA provided technical assistance. RH helped in some lab work. JB provided BRILs population. ZW supervised the research.
All authors read and approved the final manuscript.