Selective Responses And Genetic Overlaps Of The Advanced Backcrossing Introgression Lines From Directional Selection In Rice

Selection has approved its importance in both artificial and natural evolution.According to the theory of population genetics, the genetic hitchhiking effects will causethe rise in frequencies of selection-favored alleles and the related alleles at closely linkedloci. The genetic overlaps are caused by both genetic hitchhiking and pleiotropy. With theaid of SSR markers, a set of 'Advanced Backcrossing Drought Selective IntrogressionLines' (DTIL) were analyzed for population structures, especially the allelic frequenciesand pairwise linkage disequilibrium (LD). By computer simulating sampling, analysis onthe selection targeting traits and non-targeting traits were carried out, in order to evaluatethe application of allelic deviation analysis based on genetic hitchhiking theory andANOVA with selective population. Combined with ANOVA and linkage disequilibriumanalysis, the candidate loci non-targeting traits were also investigated. The genetic overlapsbetween targeting traits and the affects of directional selection on loci detection fornon-targeting traits were discussed. The main results were listed as following:1, The selective responses of the advanced backcross population under selection.Both the simulating sampling with 133 BC₂ random introgression lines (ORIGINAL)and the genotyping analysis with 343 DTILs pointed out that the main effects of directionalselection on advanced backcross population were the extreme variation of allelicfrequencies and the overwhelming numbers of highly significant pairwise LDs, which wereinfluenced by donor, recipient and the selection conditions in the real DTILs.2, The simulating selections in a real big random population.With the big random population (ORIGINAL) as control, three sub-populations(HD, PH and YLD) directionally selected according to the phenotypic data under lowlandstress and small random populations (RANDOM-â… and -â…¡) from repeated sampling, were evaluated in mapping for both selection-targeting and non-targeting traits loci allelicfrequency deviation, LD analysis and ANOVA. The results indicated that the deviationanalysis can be used in mapping the loci for selection-targeting traits. The accuracy rangedby 33.3%～57.1% depending on the traits. ANOVA with selected populations can be a wayof remedy in some loci missed by deviation analysis. The deviation noise from populationstructure and small sampling will increase the false positive rate. In mapping loci forselection non-targeting traits, highly significant pairwise LDs will increase the number anduniform the genetic effects of candidate loci detected. The results were also affected bygenetic overlaps between the targeting and non-targeting traits.3, Targeting trait analysis with Bg300 DTILs.Four sets of DTILs of two backgrounds with Bg300 as donor through selections underlowland and upland drought stresses were analyzed for yield per plant under stress. Twentyregions were mapped all over genome. Most of them belongs to donor excessive loci, andloci under reverse selection (selection favoring recipient alleles) were few and at lowsignificant levels(X²ï½ž5). The number of excessive loci detected in lowland and uplandconditions were almost equal. Comparison with the 23 loci detected for yield and paniclefertility by other researchers showed that 12 of our 24 loci were matched. Of the other locireported for the first time, especially those supported by multiple evidences such as QDt3d,QDt6b, QDt9a and QDt9c, are possibly new loci for drought tolerances.4, Milling quality and seed shapes.The analysis with 231 random introgression lines strongly indicated the geneticcorrelations underlying the trait overlaps between the milling quality and seed shapes,especially the grain length (GL). Further investigation was done in DTILs with Lemont asdonor and IR64 and Teqing as recipients. At phenotypic level the seed shapes were notaffected much by selection in DTILs, while the milling quality, especially the head rice rate(HR) were strongly influenced. The decreasing of HR in DTILs of both geneticbackgrounds indicated the negative correlation between drought tolerances and HR. Fromthe view of related loci detected, the genetic correlations between milling quality and seedshapes were still maintained in three main overlap regions for milling quality and grainshapes on chromosome 5, 6 and 7, respectively.The Association Loops (ALs) formed by highly significant pairwise LDs affected allthe milling quality and seed shape traits, by creating regions with consistent genetic effects.By the evidences from random population mapping, the possibilities of false positive are high for these regions. Influenced by ALs, the rate of accuracy for mapping loci of millingquality and seed shapes were relatively low. When the false positive loci caused by ALswere exluded, the ratio of the loci detected in the random population still functioning in DTpopulations is 100.0%, 60.0%, 33.3%, 100.0% and 50.0% for BR, MR, HR, GL and GLTR,respectively. These loci are supposed to offer clues to quality improvement for DT varieties.5, Genetic overlaps between sheath blight resistance (SBR) and drought tolerance(DT).With deviation analysis in 157 DTILs, fifty regions supported by 106 significant donorexcessive introgression cases were identified, the average chi-square value of these loci is54.3 (P=1.6×10^-12). Among 106 cases, 46 belong to the loci with supporting evidencesfrom different populations, and 4 loci were detected in more than 3 populations. Of the 50DT regions, the introgression frequencies of donor alleles were more than 0.65 in 16(32.0%) ones. With LD analysis, a complicated genetic network was discovered associatedwith drought tolerances in DTILs. It was consisted by 17 ALs, each of which includes 2-10unlinked but highly LD associated loci.Based on the same population, the candidate loci for SBR were also identified. Totallytwenty-eight SBR loci were found, of which 64.3% showed SBR improving by donoralleles. Multiple allelic diversities were found at seven SBR loci. With the comparisonthrough consensus map, 18 of the 28 SBR loci matched the SBR gene/QTL reported byother researchers. QSbr1b, QSbr1c, QSbr2a, QSbr2c, QSbr3c, QSbr6b, QSbr7b, QSbr10a,QSbr12a and QSbr12b were reported for the first time.The relationships between DT and SBR were shown at two levels. The first level is thepurged phenotype after selection. The DTILs from drought tolerance selection were betterin DT than RP. The DTILs with Teqing background were more susceptive than Teqingwhile those with IR64 background were more tolerant than RP. The second level isgenotypic level, of which the related loci can be divided into three types: the first type is thenegative associated loci, which account 33.3% of the 33 overlapped loci of SBR and DT.Of the 32 excessive loci detected in the DTILs from Teqing/Bg300, Teqing/Babaom andTeqing/Basmati, 25(78.1%) matched the SBR loci directly or indirectly. The second type ofloci associated with the improved DT and SBR consistently, which account for the remainig66.7% of the 33 overlapped loci of SBR and DT. Of the 23 excessive loci detected in theDTILs from IR64/Bg300,IR64/Babaomi and IR64/Basmati with enhanced SBR, 13(56.5%)matched SBR loci. The third type of loci affected DT and SBR independently, which include 2 SBR loci (QSbr11b and QSbr12b) and 24 DT loci.From on the above simulating and practical analysis with population from selection,some points were strongly indicated: first, the deviation analysis based on selectiveresponse can be applied in mapping loci associated with selection target with the aid ofprogeny test and ANOVA. Second, the multiple loci structure (linkage disequilibriabetween unlinked loci) will increase the number of positive loci detected and uniform theeffects of loci in mapping of non-target traits. Thirdly, the genomic regions detected formilling quality, grain shape, sheath blight resistance and drought tolerance are the possiblegenetic basis underlying the phenotypic correlations between relative traits and supposed tooffer useful information for molecular breeding.