Oryza nivara allele of a major effect QTL qFLA1.1 increases ag leaf area in rice

A stable back cross introgression line IL65 (IET22161) (Swarna/O. nivara-BC 2 F 6 ) of rice was used to map ag leaf related traits in F 2 and F 3 . A total of 12 QTLs were mapped for ag leaf related traits on two chromosomes with each QTL explaining 3 to 21% phenotypic variance (PV). Interestingly, a novel 12Mb QTL cluster (RM8094 - RM9) that controls 7 traits was identied on long arm of chromosome 1 where QTLs qSPAD1.2, qSPAD1.3 for SPAD, qFLL1.1, qFLL1.2 for ag leaf length, qFLW1.1, qFLW1.2 for ag leaf width, qFLA1.1, qFLA1.2 for ag leaf area, qPH1.1, qPH1.2 for plant height, qDTF1.2, qDTF1.3 for days to owering and qHI1.2, qHI1.3 for harvest index were co-located. Among these, one major effect QTL qFLA1.1 for ag leaf area was identied in a 9Mb region between RM8094 and RM5638. There was an adjacent minor effect QTL qFLA1.2 in a 3Mb region between RM5638 and RM9. Together these two QTLs and with leaf area increasing QTL allele from O. nivara explained 19.7% PV. The QTL for ag leaf related traits can be ne mapped and considered for breeding rice varieties with higher ag leaf area, photosynthetic rate and grain yield. length, panicle weight, primary branch number and secondary branch number. Interestingly our major effect QTL qFLA1.1 for ag leaf area, is 2.3Mb downstream to their QTLs qFLA1.1. Recent report of Wen et al. (2020) showed a major effect QTLs qTLL1 and qTLLW1 for third leaf length and third leaf width on long arm of chromosome 1. We could not nd out the positions of these QTLs because of usage of different markers. Similar study was conducted by Farooq et al. (2010) in F 2 population derived from IR64 (indica) and its derived introgression lines (from new plant rice) and identied 4 QTLs for ag leaf size. One QTL qLLnpt-1 for leaf length and qFLWnpt-1 for ag leaf width were detected in an F 2 population of HKL 69 and HFG 27 respectively. The QTL qLLnpt-1 is 11Mb downstream to our QTL qFLL1.1 and qFLWnpt-1 is on short arm of chromosome 1. The QTLs qFLLnpt-4 controls ag leaf length and qFLWnpt-4 controls ag leaf width on chromosome 4 are co-located. Tian et al. (2014) identied a locus between RM3521 and RM8111 on short arm of chromosome 1 that controls both ag leaf width and grain number per panicle in F 2 population derived from HP (indica) and Nipponbare (japonica). The markers associated with hotspot 1 are previously reported as linked to other traits also. RM8094 at hotspot1-1 is one of the most useful and informative marker within Saltol QTL for salt tolerance (Ganie et al. 2016; Chowdhury et al. 2016). RM5638 at hotspot 1–2 is linked to qbr-1.1 for brown rice yield, qkl-1.1 for kernel length, qkwt-1.1 for kernel weight, qmr-1.1 and for total milled rice (Nelson et al. 2012). But the association of hotspot 1 markers for ag leaf related traits were not reported previously.


Introduction
Rice yield is the primary target trait in breeding programs. Conventional crop improvement methods have been found limiting in breaking current yield barriers. Molecular and biotechnological tools can complement the conventional methods to enhance rice yields. Earlier, ideotype breeding was described to improve plant architecture with short plant height, more tillers and panicles with erect ag leaves to enhance rice productivity (Wang and Li 2005). In recent years, improving canopy photosynthesis is recognized as a major avenue to enhance crop growth rate, biomass and grain yield (Song et  level is also an important factor to improve crop biomass (Makino 2011). Therefore, photosynthesis was turned out as one good target for breeders to look for lines with high net photosynthetic rate to develop high yielding lines (Teng et al. 2004). In our previous study we found that the introgression lines with high net photosynthetic rate showed high yield and total dry mass (Haritha et al. 2017(Haritha et al. , 2019. In cereals, the top three leaves, particularly ag leaf, is the most essential functional organ to produce a large proportion of photo-assimilates that are later stored in grains (Peng et  ). On the other hand, larger leaves often shadow lower leaves thus reducing radiation there by decreasing the leaf photosynthetic rate. Leaf size is a complex trait and the angle, thickness, folding, leaf temperature, leaf wax and several other traits need to be optimised in a canopy for maximum radiation use and e ciency. Even breeders consider the ag leaf size as a bene cial trait for sustaining yield potential in wheat under different water-de cit conditions (Isidro et al. 2012).
The plasticity of leaf size and shape are strongly dependent on environment conditions (Tsukaya et al. 2006). Asymmetric changes in day/night temperatures, light and air humidity show strong impact on plant growth rate, leaf area biomass and dry-matter (Stuerz and Asch 2019). High night temperature increases leaf growth including leaf area and photosynthesis (Jing et al. 2016). Leaf blade expansion is inhibited and petiole length enhanced during low light conditions in Arabidopsis (Kozuka et al. 2005). However, many studies show that these leaf morphological traits are signi cantly correlated with yield related traits. For example, ag leaf length was positively correlated with days to Several QTLs were reported for ag leaf size and shape ( ag leaf length, width, and area) using diverse mapping populations such as  Kobayashi et al. (2003) identi ed nine genomic regions affecting ag leaf development in 190 RIL population derived from rice varieties Milyang23 and Ashikari. Among these, three chromosomal regions (group I) had stable QTLs that increased both FLL and FLW, whereas four regions (group II) had QTLs only for increasing FLL and other two regions (group III) had only QTLs for increasing FLW. Four QTLs were identi ed in four F 2 populations derived from IR64 and its introgression lines (IR64/new plant type rice) for ag leaf length and ag leaf width on chromosomes 1, 2 and 4. Among these qFLLnpt-4 at RM3843 and qFLWnpt-4 at RM17483 on chromosome 11 led to longer and wider ag leaves in IR64 introgression lines HFG39 population when allele was from YTK298 genome (Farooq et al. 2010). Sonah et al. (2012) identi ed a stable and major QTL qLL12.1 for ag leaf length between RM247 and RM6296 across 3 different climatic zones of India. Similarly, Cai et al. (2015) identi ed 30 QTLs for ag leaf length, width and area in 2 DH populations of ZYQ8/JX17 and CJ06/TN1. Among these qFLL-4b for ag leaf length between RM252 and RM3276 on chromosome 4, qLW-12 for ag leaf width between G148 and RG413 on chromosome 12 and qFLA-2a for ag leaf area between CT87 and G1234 on chromosome 2 showed high PV and additive effects (26.3, 6.98), (20.8, 0.12) 24 QTLs for leaf morphology were identi ed in a RIL population derived from a cross between japonica cultivar Rekuangeng (RKG) and indica cultivar Taizhong1 (TN1). Among them 8 QTLs were detected for leaf length and 16 QTLs for leaf width on chromosome 1 (Wen et al. 2020). The lines used in these studies were developed from indica x indica, indica x japonica and japonica x japonica crosses. Flag leaf related QTLs have not yet been reported from wild species or their derivatives in rice. In all, 39 QTLs for ag leaf area have been reported using different mapping populations (www.gramene.org accessed on 20th November 2019).
Oryza is an agronomically important genus containing species with highly diverse morphological characteristics; however, the major genetic variations have not yet been fully exploited in rice breeding (Wambugu et al. 2013;Govindaraj et al. 2015;Haritha et al. 2016Haritha et al. , 2018a. Therefore, wild species could be utilized as a potential source for further improvement in leaf structural and physiological traits which ultimately improve the e ciency of resource capture in modern cultivars equivalent to expected photosynthetic use e ciency in C4 rice. All these studies on FLL, FLW and FLA indicate that these traits are strongly related to grain yield. However, all previous QTL mapping studies for ag leaf were conducted in populations derived from intra-speci c crosses of Oryza sativa. QTLs for ag leaf related traits have not been mapped previously using a wild species derived line as a parent. The aim of present study was to locate major QTLs for ag leaf related traits ( ag leaf length, width and area) and its association with grain yield and related traits using simple sequence repeat markers. IL65 was also tolerant to prolonged shade (Panigrahy et al. 2018) and heat (Prasanth et al. 2017) and also showed high photosynthetic rate (Rao et al. 2018). The salient features of parents are given in Table S1 in Online Resource 2. Since, Swarna and IL65 have differences in many agro-morphological traits, culm and ag leaf anatomical analysis also performed at owering stage. The scheme for development of mapping populations and con rmation of true F 1 s using SSRs are shown in Fig. S1 in Online Resource 1.

Phenotyping
The quantitative data of 12 morphological and yield related traits were measured in 473 F 2 plants and of 21 traits for three plants each of 427 F 3 families in three replicates. Flag leaf size related traits were measured at heading stage and other yield related traits at maturity stage.

Measurement of ag leaf related traits
In F 2 single plant data were taken for all 12 traits, but in F 3 families the mean of middle three plants of centre row were taken from each replication. The relative content of leaf chlorophyll was measured by taking three SPAD measurements per ag leaf of middle three plants using Minolta 502 chlorophyll meter. Flag leaf length (FLL) was measured from leaf base to tip in 3 fully expanded ag leaves of main stem at heading and expressed in cm. Flag leaf width (FLW) was measured at the middle of the ag leaf where it is widest at heading and expressed as cm. Flag leaf area (FLA) was calculated as FLA = FLL x FLW x A and expressed in cm 2 (where A = 0.747, a constant value (Stickler et al. 1961).
These 3 traits FLL, FLW and FLA determine the fag leaf size. Thickness (mm) of 3 ag leaves from 3 plants was measured using vernier callipers.

Measurement of grain yield and yield related traits
Days to owering (DTF) in F 2 was determined by counting number of days taken from the day of sowing to rst panicle emergence in each plant, and as days to 50% owering (DFF) when 50 percent of the F 3 plants in a family owered. PH (cm) was measured from the base of stem to tip of main panicle at maturity. Number of tillers (NT) and number of panicle bearing tillers/productive tillers (NPT) were counted manually at maturity. After harvesting grain yield (YLDP) was measured by taking the mean weight of dried (12-14% moisture) grains from three plants of each replication and the mean yield was expressed in grams (g). Similarly, the weight of above ground biomass (BM) of three well dried plants (on which grain yield was taken) was measured and average was taken and expressed in grams (g). Total dry mass (TDM) was calculated as TDM = YLDP + BM. Similarly harvest index (HI) was calculated as the ratio between YLDP and TDM of the plant and expressed in percentage (HI = YLDP/ TDM x 100). Panicle length (PL) was measured from base (including peduncle length) to the tip of 3 panicles in each replication and expressed in cm. Number of primary branches (NPB) and number of secondary branches per panicle whose length was measured was counted and average was taken. Number of lled grains per panicle (NFG), un lled grains per panicle (UFG), total grains per panicle (GNP) including un lled grains, was counted manually from 3 panicles of each replication. Spikelet fertility (SF) was calculated as ratio between NFG and GNP and expressed in percentage (SF = NFG/ GNP x 100). Panicle weight (PWT) of the main 3 panicles was measured and expressed in grams (g). Finally, 1000 grain weight (TGW) expressed in grams (g) was taken from 1000 randomly selected dried grains from each replication.

Statistical analysis
The mean data for each trait in F 2 and three replications of F 3 was subjected to statistical analysis. Analysis of variance (ANOVA) and multiple correlations were performed using STAR ver 2.0.1 (http://bbi.irri.org/products). Descriptive statistics and signi cance of variance components were determined using PB-Tools ver 1.4 (http://bbi.irri.org/products). Broad sense heritability was estimated using the method of Johanson et al. (1955) as heritability = [Genotypic variance/Phenotypic variance] x 100 and expressed in percentage.

Genotyping of mapping population
The leaves of 427 F 2 plants and parents were collected at 60 days old plant stage. Genomic DNA was isolated using a modi ed protocol of Zheng et al. (1995). The quality and quantity of DNA was measured through spectrophotometry using Nanodrop (ND 1000, Thermo Scienti c, Madison, USA). DNA was diluted with TE (Tris EDTA) buffer to make the nal concentration to 50ηg/µl for PCR analysis. A total of 1609 RM (rice microsatellite) primers from all 12 chromosomes were used to survey the polymorphism between parents Swarna and IL65. The primers were chosen based on their genome-wide distribution and earlier reports (McCouch et al. 2002). It may be noted that IL65 is an introgression line (BC 2 F 6 ) and genotypic data showed IL65 has loci which are either homozygous for Swarna (AA), or O. nivara (BB), or heterozygous loci (AB) with alleles from both the parents. Only markers which were of different sizes and were homozygous in each parent [designated AA in Swarna and BB in IL65] were considered polymorphic for further segregation analysis in F 2 . Presence absence variation and non parental bands were not considered as polymorphic. Any marker polymorphic between Swarna and IL65 was assumed to be due to introgression from O. nivara IRGC81848 in IL65. Of these 110 markers were used to analyze marker segregation in F 2 population.
The PCR reaction for simple sequence repeats (SSR) was performed in 10µl reaction volume containing 50ng of template DNA, 0.2µM of each primer (both forward and reverse primers) and Emerald Amp PCR Master Mix (Takara Bio USA, Inc.). The PCR ampli cation was performed under the following conditions: initial denaturation at 94 0 C for 5 min, followed by 35 cycles of denaturation at 94 0 C for 30s, annealing at 55 0 C for 30s, extension at 72 0 C for 1 min, followed by the nal extension at 72 0 C for 7 min. Following ampli cation, the products were resolved on 3% agarose gel and bands were scored according to segregation of parental alleles in all samples for each of the primers separately. . The set of co-factors was adjusted if the most likely position of the QTL differed from that identi ed in the co-factor selection round, and subsequent rounds of MQM mapping were performed. Markers were removed as a co-factor if their LOD value dropped below the signi cance threshold. When LOD values in other regions reached a signi cant level, the MQM was repeated by adding new markers as co-factors until a stable LOD pro le was reached. The con dence interval for each QTL was set at the 1-LOD support interval. This corresponds approximately to a probability of < 0.05 for declaring false positives in the entire genome based on sparse-map model (Lander and Botstein 1989). Adjacent QTLs on the same chromosome were considered as different when the curve had a minimum between peaks that were at least 1-LOD unit below either peak or when the support intervals were non-overlapping (with at least 20 cM) (Lander and Botstein 1989). The LOD (logarithm of odds) value above 2.5 was considered for signi cant QTL detection.
The position of the QTL was estimated as the point of maximum LOD value in the region under consideration. The phenotypic variance explained by a single QTL was calculated as the square of the partial correlation coe cient (R 2 ) with the observed variable, adjusted for cofactors. The additive effect of a putative QTL was estimated as half the difference between two homozygous classes.
Epistatic interactions between different marker loci were determined for all traits in F 2 and F 3 populations for single environment at a LOD of 5.0 using ICIM-EPI function in ICiMapping ver 4.1 (Meng et al. 2015).

Identi cation of CSSLs
The genotypic data of 93 polymorphic markers in 427 F 2 population was used in CSSL nder ver1.4 (http://mapdisto.free.fr/CSSLFinder/), to identify a minimal set of chromosome segment substitution lines (CSSLs) in Swarna background that represents the entire introgressions from O. nivara (i.e. derived from IL65).
Identi cation of putative candidate genes within the major QTL The candidate genes were searched within the hotspot 1 (12Mb) that controls ag leaf size and yield related traits were determined by using gramene data base (http://www.gramene.org/) and RAP-DB (https://rapdb.dna.affrc.go.jp/).

Phenotypic variations of parents
The phenotypic differences between Swarna and IL65 were signi cant for all traits, except for number of lled grains per panicle (NFG), intrinsic water use e ciency (P N /g s ) and ratio of chlorophyll a to chlorophyll b (Chl a/b). IL65 showed signi cantly higher ag leaf area, plant height, days to 50% owering, biomass and total dry mass than Swarna (Fig. 1a, b, c, d, e, f, g). The mean, standard deviation, skewness and kurtosis for all traits are shown in Table S2 in Online Resource 2. Fifteen out of 33 traits were negatively skewed with values from − 1 to + 1. The transverse sections of culm showed more number of vascular bundles in IL65 (34) compared to Swarna (30) and large air spaces below epidermis only in Swarna absence in IL65 (Fig. 1h, i). Flag leaf showed laments of cells found only in midrib of Swarna, absent in IL65 (Fig. 1j).

Phenotypic variations of F 2:3 populations
Transgressive segregants were observed for all traits in F 2 and F 3 (Fig. S3, Online Resource 1). Analysis of variance (ANOVA) showed signi cant (P < 0.01) variations for all traits in F 3 population ( Table 1). Most of the traits showed near normal distribution. But, ag leaf length, days to owering, plant height and harvest index in F 2 , and ag leaf thickness, plant height, days to 50% owering, spikelet fertility, panicle length, thousand grain weight and harvest index in F 3 showed negative skewness ( Fig. S4 and Fig. S5, Online Resource 1). Flag leaf length ranged from 11 to 40cm, ag leaf width from 0.5 to 5cm and ag leaf area from 5.98 to 63.5cm 2 . The coe cient of variation for ag leaf length, ag leaf width and ag leaf area ranged from 10.6 to 26.7 in F 2 and 18.5 to 30.8 in F 3 . The CV was highest for yield per plant (64% in F2 and 52.4% in F3) and lowest for days to owering (4.2% in F 2 and 4% in F 3 ). (Table S3 and Table S4, Online Resource 2). whereas the correlation with FLW, NFG and NSB was signi cant at P < 0.01 level, and with PL, NPB and YLDP was less signi cant (P < 0.05) compared with other traits (Table 3). In addition, YLDP showed highly signi cant (P < 0.005) positive correlation with BM, TDM, HI, NSB, NFG, GNP, SF and PWT, but it was less signi cant for PL, NPB and TGW (P < 0.05). YLDP correlated signi cantly (P < 0.05) with FLL and FLA.

Major effect QTL alleles
Only two major effect QTL alleles were detected in F 2 and F 3 populations (Fig. 2b). One QTL qDTF1.2 (RM8094-RM5638) for days to owering was detected at LOD 6.7 in F 2 . The trait enhancing QTL allele was from Swarna and explained 13.5% PV. The other adjacent QTL qDTF1.3 (RM5638-RM9) for days to owering was detected at the highest LOD of 7.4 and explained 9.1% PV. Together these two QTLs explained 22.6% of PV. In F 3 , one major effect QTL qFLA1.1 (RM8094-RM5638) was detected for ag leaf area and explained PV of 21%. The trait enhancing QTL allele was from O. nivara and had an additive effect of 2.74cm 2 . The adjacent QTL qFLA1.2 was also detected at a high LOD of 7.3 with a PV of 8% and together these two QTLs explained 29% of PV.

Co-localization of QTLs
The QTLs identi ed for different traits often clustered in the same chromosomal regions. Two chromosomal regions with three or more QTLs for different morphological, physiological and yield related traits were identi ed on two chromosomes. On chromosome 1 two contiguous QTL clusters (RM8094-RM5638 and RM5638-RM9) were detected for seven traits and three contiguous QTL clusters on chromosome 2 (RM207-RM3774, RM3774-RM13260 and RM13260-RM5460) for nine traits (Fig. 2c).
The comparison of QTLs on chromosome 2 revealed that all the yield related traits whose loci are on chromosome 2 are also highly correlated. For example, number of tillers, number of productive tillers, yield per plant, biomass and total dry mass were signi cantly correlated and also co-localized in the region between RM207 and RM3774 in F 2 . Likewise, plant height, panicle length and grain number per panicle are highly correlated and their QTLs co-localized in the region between RM13260 and RM5460 in F 3 (Table S12, Online Resource 2).

Epistatic interactions for ag leaf size and yield related traits
Epistatic interactions (digenic) between the marker loci for single environment detected 167 interactions for all 10 traits except for DTF and SPAD in F 2 . Two signi cant digenic interactions were detected for FLL between the marker loci on chromosome 4 and chromosome 6. Thirteen interactions were detected for FLW and eighteen interactions were detected for FLA on chromosome seven and 11 simultaneously. Chromosome 11 showed high number of (5) interactions for FLA with the marker loci on other chromosomes and showed high LOD of 8.9.
SPAD, PH, NT, NPT, BM, TDM and DTF showed lowest interactions and YLDP and HI showed highest epistatic interactions across all the chromosomes (Fig. S8, Online Resource 1). Interestingly, the loci on chromosome 2 interacted with QTLs on different linkage groups. In F 3 , 496 signi cant epistatic interactions were detected for all 21 traits except for SPAD. Among these, FLL showed one epistatic interaction between the marker loci on chromosome 2 (RM5404-RM12924, 30cM) and chromosome 11 (RM332-RM209, 25cM) contributing 7.4% of PV at a LOD of 5.9.
Five signi cant interactions were detected for FLW on chromosome 1, 2, 4 and 6. Similarly, FLA showed three digenic interactions on chromosome 1, 6 and 12. Of these, the interaction between the marker loci on chromosome 1 (RM1-RM579, 20cM and RM8094-RM5638, 95cM) contributes high percentage of PV 8.6 with a LOD of 5.1 (Fig. 3). YLDP, BM, TDM, HI, NFG and GNP showed lowest interactions and DFF showed highest interactions followed by PL, NPB, PH, SF and TGW across all chromosomes (Fig. S9 (Table S13, Online Resource 2). These elite CSSLs are useful genetic resource for ne mapping.
In silico analysis of candidate genes within the major effect QTL The physical length of major effect QTL qFLA1.1 for ag leaf area between RM8094-RM5638 (hotspot1-1) is 9Mb and RM5638-RM9 is 3Mb (hotspot1-2) (together 12Mb) on long arm of chromosome 1. It contains 1063 putative candidate genes (hotspot 1-1 contains 801 and hotspot 1-2 contains 262), which maybe involved in improving ag leaf and yield-related traits. Genes already reported for ag leaf size and yieldrelated traits were found in this region. These reported putative candidate genes are listed in Table S14, Online Resource 2.

Discussion
Flag leaf is the main photosynthetic organ and plays a pivotal role in capturing of light, and energy utilization at grain lling stage. Thus ag leaf area has a direct impact on photosynthesis e ciency and increasing leaf area can help increase yield. Related wild species have the potential to increase ag leaf area. We report here that an Oryza nivara allele of a major effect novel 9Mb QTL qFLA1.1 increases ag leaf area in rice. In the present study two hotspot pleiotropic QTL or QTL clusters were identi ed each on chromosome 1 and 2. The hotspot 1 (RM8094-RM5638-RM9) region on long arm of chromosome 1 is novel and not reported earlier for ag leaf and yield-related traits in rice using wild derived introgression lines. This hotspot 1 region has a cluster of 11 QTLs for 7 traits ( ag leaf length, ag leaf width, ag leaf area, SPAD, days to owering, plant height and harvest index) mapped in F 2 and F 3 . In F 2 , one QTL qDTF1.2 for days to owering was located between RM8094 and RM5638 (hotspot 1-1) on chromosome 1 and another QTL qDTF1.3 for days to owering was located adjacently between RM5638 and RM9 (hotspot 1-2). The trait-enhancing QTL allele was from Swarna in both.
Heading date and days to fty percent owering are key determinants of rice maturity and yield and in uenced by many environmental factors such as day length, temperature, light intensity and nutrients. In the present study ten QTLs were identi ed for days to owering in F 2 , three QTLs were located on chromosome 1, one on chromosome 3, two each on chromosome 6, 10  was located 77.6Kb upstream to Ghd7, a major QTL reported previously for delayed owering, increased plant height, grain number per panicle and grain yield under long day duration (Xue et al. 2008;Yan et al. 2011Yan et al. , 2013. In F 3 , the only one major effect QTL detected was qFLA1.1 for ag leaf area and located between RM8094 and RM5638 (hotspot1-1) on chromosome 1 with a LOD value of 3, PV of 12.7%, and a high additive effect of 2.45cm 2 . There was an adjacent minor QTL qFLA1.2 for ag leaf area between RM5638 and RM9 (hotspot1-2) with high LOD of 6.4 and 7% PV and additive effect of 1.66cm 2 . These two QTLs had trait enhancing alleles from O. nivara. The leaf morphological and physiological traits were controlled by many QTLs and were in uenced by several environmental factors (Zhang et al. 2009;Haritha et al. 2018a).
In present study ag leaf length, width and area are mutually correlated. The correlation between ag leaf size and yield-related traits revealed ag leaf length and area were signi cantly (P previously reported as linked to other traits also. RM8094 at hotspot1-1 is one of the most useful and informative marker within Saltol QTL for salt tolerance (Ganie et al. 2016;Chowdhury et al. 2016). RM5638 at hotspot 1-2 is linked to qbr-1.1 for brown rice yield, qkl-1.1 for kernel length, qkwt-1.1 for kernel weight, qmr-1.1 and for total milled rice (Nelson et al. 2012). But the association of hotspot 1 markers for ag leaf related traits were not reported previously.
The other QTL cluster was hotspot 2 (RM207-RM3774-RM13260-RM5460) that has 11 minor effect QTLs for 9 traits were detected on long arm of chromosome 2. However hotspot 2 − 1 (RM207-RM3774) and hotspot 2-2 (RM3774-RM13260) in uences number of tillers, number of productive tillers per plant, yield per plant, biomass and total dry mass, whereas hotspot 2-3 (RM13260-RM5460) in uences panicle length, number of secondary branches, number of lled grains and grain number per panicle in F 2 and F 3 . The QTLs for source related traits were colocated with some sink-related traits (Cui et al. 2003;Zhang et al. 2015;Wang et al. 2020). In our study 5 minor effect QTLs for SPAD, FLL, FLW, PH and HI were also co-located in hotspot1 (RM8094-RM5638-RM9) region on chromosome 1. This is a novel hotspot region detected for these traits in present study.
The QTL clusters represent the genes with either pleiotropic effects on many traits or close linkage of different genes in rice (Wang et al. 2012) and wheat  Hence these are worthy of not only ne mapping and cloning for gene discovery but also may be useful for use in MAS or MAB as we show the effect of alleles from O. nivara is consistent across generations and populations.
Flag leaf traits and yield were collocated on chromosome 1 and chromosome 2 in our study (hotspot 1 and hotspot 2) and this has been reported previously also. Tang et al. (2018) identi ed 14 QTLs for ag leaf length and 9 QTLs for ag leaf width in a CSSL population of Zhenshan 97 (indica) and Nipponbare (japonica). Among these, two QTLs (qFL7-2 and qFW7-2) were detected in the same region near the heading date QTL GHD7.1 and explained 11 and 5.5% PV. Further, validation of this yield related region which controls ag leaf length, ag leaf width, photosynthetic capacity, owering time and yield potential is required. 43 QTLs for ag leaf size, shape and yield related traits were identi ed from 135 RILs derived from 93-11(indica) and cv. Peiai64s (javanica) (PA64s) and a major and novel QTL qFLW7.2, for ag leaf width was identi ed between INDEL7-2 and INDEL7-3 on chromosome 7. Interestingly qPY7 for plant yield and qFLW7 were co-located within the same interval region on chromosome 7 . Li et al. (1999) identi ed ve QTLs related to ag leaf area on chromosomes 2, 5, 6, 7 and 9 in F 2 population derived from Lemont (japonica) and Teqing (indica). They concluded that leaf area was positively correlated to grain yield, and QTLs in uencing ag leaf related traits and grain yield-related traits were mapped to similar genomic regions and showed positive in uence on the traits. Thus, in addition to chromosome 1 other chromosomes also have ag leaf area QTLs collocated with yield trait QTLs. Since the QTL on chromosome 1 is a major QTL with increasing effect from the related wild species O. nivara and identi ed in a largely Swarna background, it is reasonable to assume that once ne mapped and cloned it holds promise for use in marker assisted transfer. It may be noted that NAL1 for narrow leaf has recently been cloned (Qi et al. 2008).
The O. nivara derived CSSLs were reported in BC 2 F 8 but they were not screened for leaf traits (Surapaneni et al. 2017, Balakrishnan et al. 2020).
Any CSSL with signi cantly higher or lower ag leaf area than Swarna can be used for narrowing the 9Mb QTL further. CSSLs are excellent genetic resource to understand the genetic architecture of complex traits. In present study 55 CSSLs were identi ed from Swarna and O. nivara derived IL65 population using 93 SSRs. Among these three CSSLs, CSSL 41 showed signi cantly higher SPAD, plant height, CSSL 57 for plant height, and CSSL 325 for number of lled grains and grain number per panicle were signi cantly higher than Swarna and IL65. These elite In present study a novel chromosomal regions was identi ed for SPAD (qSPAD1.2, qSPAD1.3), ag leaf length (qFLL1.1, qFLL1.2), ag leaf width (qFLW1.2, qFLW1.3), ag leaf area (qFLA1.1, qFLA1.2), plant height (qPH1.1, qPH1.2), days to owering (qDTF1.2, qDTF1.3) and harvest index (qHI1.2, qHI1.3). All these QTLs were collocated within the 12Mb region between the intervals of RM8094-RM5638 (9Mb) and RM5638-RM9 (3Mb). This region contains 1063 genes. Of these, there were six reported candidate genes three in each region RING-nger E3 ligase, that interacts with the gene IPA1. This is located within hotspot 1-1 region on chromosome 1 of our study, where several minor QTLs for ag leaf and yield-related traits were co-located. Wang et al. (2011) ne mapped the major effect QTL qFL1 that controls ag leaf size to a candidate gene OsFTL1. OsFTL1 showed pleiotropic effect on ag leaf size, heading date and other yield related traits. It is interesting to note that our QTL qFLA1.1 for ag leaf area identi ed in F 3 is the same region where the major QTL qDTF1.2 for days to owering was identi ed in F 2 of our study. Thus, the QTLs in hotspot 1 region are high priority regions and it is worthy for further ne mapping to identify the causal genes.

Conclusion
The QTLs in uencing ag leaf size related traits and several grain yield related traits were mapped to same genomic regions and showed positive in uence on the traits. Therefore it is possible to improve grain yield by genetic improvement of ag leaf length, ag leaf width and ag leaf area with the aid of molecular markers. The major effect QTL qFLA1.1 for ag leaf area in hotspot 1 region is novel and in 9Mb region QTL cluster for correlated yield related traits. It is worthy of ne mapping and functional validation of markers for use in marker assisted selection for ag leaf size related traits use in MAS. Five ILs showed higher grain yield (g) FLL, FLW, PH, BM and TDM than both parents  Table   Table 3 is only available as a download in the Supplemental Files section. Figure 2 a Chromosome-wise arrangement of QTLs from F 2 and F 3 populations. b major QTLs (qDTF1.1) identi ed for days to owering in F 2 and (qFLA1.1) for ag leaf area in F 3 population. c co-localization of QTLs for different traits on chromosome1 and 2. Chr denotes chromosome, *denotes QTLs from F 2 , # denotes QTLs from F 3.

Figure 3
Cyclic representation of epistatic QTLs. a F 2 -ag leaf length. b F 2 -ag leaf width. c F 2 -ag leaf area. d F 3 -ag leaf length. e F 3 -ag leaf width. f F 3 -ag leaf area Figure 4