Dissection,Fine-mapping And Detection Power Analysis Of A Minor QTL,qTGW1.1 For 1000-grain Weight In Rice

Most agronomical traits of crop species are complex traits controlled by multiple genes and affected by environmental factors.While major QTLs for yield-related traits have long been targeted for fine-mapping and cloning in rice,minor QTLs have yet been paid to the attention until recently.Using NILs derived from rice indica cross Zhenshan 97?ZS97?///ZS97//ZS97/Milyang 46?MY46?,we previously dissected four minor QTLs for grain weight in a 12-Mb region on the long arm of rice chromosome 1.One of the QTLs qTGW1.1 was delimited within a 376.9-kb interval flanked by DNA markers Wn28382 and RM11554.In the present study,this QTL was targeted for dissection,validation and fine-mapping.During the mapping process,we found that the statistical power is low for the minor QTLs.Therefore,the qTGW1.1a dissected in this study was characterized using NIL populations in order to explore measures of increasing detection power.The main results are as following: 1 Dissection,validation and fine-mapping of qTGW1.1Four plants were identified from the BC₂F₁₀ of ZS973/MY46,which carried the heterozygous regions that overlapped and covered the whole region of qTGW1.1.Four sets of NILs?BC2F11:12?named as Z1,Z2,Z3 and Z4 were developed from the selfing progenies of the four plants,and were grown in two locations having distinct ecological conditions.QTL analysis showed that two QTLs for grain weight were included in the qTGW1.1 region.Following the nomenclature of qTGW1.1,they were designated as qTGW1.1a and qTGW1.1b.The qTGW1.1a located within the 120.4-kb interval Wn28447-RM11543 significantly affecting all the three traits with the enhancing allele from ZS97,showing a stronger influence on grain weight than on grain length and width.The qTGW1.1b located in the 521.8-kb interval RM11554-RM11569 significantly affecting grain weight and length with enhacing allele derived from MY46,having a stronger influence on grain length than on grain weight.Afterwards,five NIL-F2 populations were developed from the selfing progenies of five BC2F13 plants with the heterozygous region covering qTGW1.1a or qTGW1.1b.They were grown at Hangzhou,Zhejiang in the summer of 2015.Two of the populations detected significant allelic effects with the ZS97 allele increasing grain weight by 0.14 g-0.19 g,the phenotypic variation explained ranged from 4.15% to 6.89%.The other three populations detected significant QTL effects for grain weight or length with the MY46 allele increasing grain weight by 0.11 g-0.12 g,grain length by 0.19 mm-0.021 mm,the phenotypic variation explained ranged from 3.19% to 5.91% and 4.04%-5.38%.These results further validate the effects of qTGW1.1a and qTGW1.1b.To further validate and narrow down qTGW1.1a to a smaller interval,three sets of BC2F13:14 NILs were constructed with the overlapped segregating regions covering only qTGW1.1a but excluding qTGW1.1b.The three populations were grown at Hangzhou,Zhejiang in the summer of 2015.Significant QTL effects with the enhancing allele from ZS97 were detected for grain weight in the three populatins,the additive effect was 0.07 g to 0.08 g,phenotypic variation explained was 5.85% to 9.91%.Thus qTGW1.1a could be delimited in the common region Wn28382-Wn28497 of the three NILs.In addition,qTGW1.1a was located between markers Wn28447 and RM11543 using the BC2F11:12 NILs.Thus qTGW1.1a was fine-mapped in a 49.4 kb region of Wn28447-Wn28497 by comparing the QTL interval of the two trials.This region contains six predicated genes.One of the genes LOC_Os01g49470 encoding E3 ubiquitin ligase is probably the candidate gene,because the ubiquitin-proteasome pathway have been reported to play a very important role in regulating rice seed size.2 Detection power analysis of qTGW1.1aZS97 and MY46 homozygous genotypic lines with extremely high and low grain weight were selected from a set of NILs(BC₂F₁₀:11)that showed the effects of qTGW1.1a in previous study.Six NIL-F2 populations were developed through hybridization between the selected lines.According to the grain weight of parental lines,the six populations could be divided into three groups,which were low ZS97 × low MY46,low ZS97 × high MY46,and high ZS97 × high MY46.The mean values and range of 1000-grain weight were similar among the six populations,and no obvious difference among groups was observed,indicating that there was no allele difference for grain weight QTLs in the genetic background among the parental lines,and the variation of 1000-grain weight among different lines within the same genotype of the NIL set used in previous study is caused by random error.QTL analysis was performed for the six NIL-F2 populations.Among the six populations,significant QTL effects for 1000-grain weight,grain length and width were detected in three,two and one populations,respectively.Thus,the actual statistical power was 50.0%,33.3% and 16.7% for the corresponding traits,indicating that the statistical power of the minor QTL qTGW1.1a is low.Theoretical statistical power was also analyzed using G*Power 3.1 based on the mean values of genotypic effect,standard deviation within the same genotype and sample size of the six populations,the estimated values were 65.8%,62.1% and 13.3% for grain weight,grain length and width,respectively.The magnitude of actual power was similar with that of theoretical power for the three traits,but slightly lower values were observed as well.Thus,it is suggested that slightly larger sample size than that of theoretical estimation should be applied in experimental design.A BC₂F₁₀ plant heterozygous in qTGW1.1a region was identified from the progenies of the original BC₂F₁₀ heterozygote.This plant was selfed and advanced to the BC2F11 generation,from which two ZS97 and two MY46 plants homozygous in the target region were selected.The selfing seeds of the four plants constituted a set of NILs?BC2F11:12?.The NILs was grown for two replications,and ten plots for each replication.The four lines were randomly arranged in each plot.ZS97 genotypic lines showed relatively higher phenotypic values than that of MY46 genotypic lines in eight,eight and seven plots for 1000-grain weight,grain length and width.This consistent phenotypic performance between two genotypic lines within different plots suggests that adjacently arrangment of two genotypic lines could increase the detection ability of minor QTLs.Each component of phenotypic variation of the NIL set was further analyzed.The variation among different plots within the same lines is actually due to the variation of the same materials planted in different positions,which directly reflects the random error.This variation explained 36.89%,48.02% and 42.10% of phenotypic variation for 1000-grain weight,grain weight and width,respectively.In the mean time,the values of random error were 43.68%,37.69% and 35.79% for the corresponding traits.The two variations explained comparable and commonly accounted most proportion of phenotypic variation.In contrast,the variation between two genotypes explained 5.65%,11.71% and 6.70% phenotypic variation for 1000-grain weight,grain length and width,respectively,which is much less than that of random error.These results suggest that minor QTLs as qTGW1.1a is difficult to show the bimodal distribution even in NIL background.