After 8 weeks, the seeds in the bags containing the tags were collected and stored for 1 month to break dormancy.
The seeds were planted in trays and kept in a screen house. Phenotypic trials and evaluations were conducted over two successive growing seasons, — and — using an alpha lattice design with incomplete blocks replicated twice. The roots were harvested 12 MAP and chopped into equal slices 5 cm using a fabricated machine cutter. Seven roots per plant and a maximum of seven slices per root were chosen at random for disease scoring.
Genomic DNA was extracted from fresh young leaf tissues harvested from the F 1 plants following a modified method of Dellaporta Dellaporta et al. Genotyping by sequencing GBS Elshire et al. SNPs were named according to the chromosome number Roman for v 5. A one-step map approach is the most common strategy for linkage mapping in cassava Sraphet et al. Goodness of fit between observed and expected segregation ratios was evaluated with the Chi squared test.
Markers were grouped using the regression method at a minimum LOD threshold of 5. The recombination frequencies were converted into map distances centiMorgans using the Kosambi mapping function. The position of the markers in each linkage group were obtained by considering their contribution to the average goodness of fit mean Chi square and the nearest neighbor fit N. Fit value. Based on these criteria, markers were removed or added to each linkage group and calculations redone until the best fit and order was obtained.
Summary statistics for the phenotypic data were calculated using GenStat Ripatti et al. The mean, skewness, kurtosis, and Shapiro-Wilk normality test were used to infer the distribution and normality of the data.
Box plots and normal plots Q—Q plots were used to inspect the quality of the data and identify outliers. The mean of each genotype across the replicates in each year and site were calculated and used for QTL mapping. A genome wide LOD threshold with P value of 0. Additive a and dominance d effects were calculated based on the method of Muchero Muchero et al.
According to Muchero Muchero et al. ICIM used in this study is an improved algorithm of composite interval mapping suitable for biparental crosses Zhang et al. It has increased detection power, a reduced false detection rate, and less biased estimates of QTL effects. This approach minimizes the bias due to Beavis effect Xu, associated with QTL analysis using a small population size and was suitable for our population that comprised of individuals.
The main purpose of genome-wide re-sequencing was to investigate whether there are any genomic regions in Kiroba derived from M. The varieties for this study were; M. DNA extractions, library preparations and sequencing was done as described in Bredeson et al. The first 10 bases were trimmed using fastx trimmer and then de novo assembly performed using abyss-pe Simpson et al.
Default parameters were used with a k-mer of The purpose of assembling the Kiroba genome was to obtain high quality scaffolds for alignment and SNP analysis. The quality of the Kiroba assembly was assessed by N50 length statistics derived from the abyss-pe output. Based on the results of the assembly, scaffolds and contigs smaller than bp were discarded to avoid using low quality reads in subsequent analysis.
Assemblies of M. For each SNP, the absolute value for difference between the M. A loop was created to look at 1, bp at a time starting at position 1. Chromosomal locations of introgression regions found in Kiroba were compared to those found in other genotypes by incorporating Kiroba into an earlier analysis Bredeson et al.
Comparative genome analysis was done to investigate whether Kiroba shared the same M. Kiroba is thought to be a possible former Amani interspecific hybrid just like Namikonga and is postulated to have found its way into farmer fields and now being grown with unknown identity in Tanzania. Putative introgression regions were aligned to the cassava reference genome v6 Bredeson et al.
The gene list for genes within the QTL regions and introgression segments were analyzed and tested for significance based on P -value and FDR. Genetic relatedness analysis between Kiroba and 40 other accessions Goodstein et al. The SNP data for all the accessions was obtained from Phytozome v. A network plot showing the first degree relatedness was drawn using Cytoscape Shannon et al.
Controlled crossing through 1, hand pollinations between Kiroba and AR resulted in 2, seeds. The mapping population was validated by SSR screening to identify off-types and true crosses. The 15 F 1 progeny with unexpected alleles, presumably having received pollen from elsewhere were regarded as off-types and those with the expected allelic composition were classified as true F 1. The integrity of the mapping population was further tested by cluster analysis of the GBS data based on identity by state IBS Figure 1.
The parent—offspring comparisons are denoted by red dots. Based on this analysis of GBS data, a further 75 off-types were identified as well as true progeny. All 90 off-types were excluded from downstream analysis.
Figure 1. In the figure, parent—parent comparisons are represented by a blue circle, parent—offspring comparisons by red dots and offspring—offspring comparisons by black dots. Summary statistics of phenotypic data obtained for two growing seasons — year 1 and — year 2 in two locations, Chambezi C1 and C2 and Naliendele N1 and N2 are presented in Table 1.
The highest mean 3. The standard error of the mean SE ranged from 0. Root necrosis was also positively skewed in N1 and N2 skewness 2. CGM was also positively skewed at C2. The highest variance was obtained for root necrosis in C1 1. The SE followed a similar trend as the variance with the highest value reported for necrosis at CI and lowest 0. Table 1. Descriptive statistics for phenotypic data for two years — 1 and — 2 for two sites, Chambezi C and Naliendele N in Tanzania.
The final linkage map Figure 2 is composed of F 1 individuals with 1, SNP markers distributed across 21 linkage groups and spanning 1, cM. Figure 2. High-density genetic linkage map based on SNP markers. They have maximum LODs of 2. Seven putative QTL were also identified, five of these were associated with both CBSD root necrosis and foliar symptoms on chromosomes 4, 6 and 11 Table 4. Table 3. Figure 3. Table 4. The rest of the QTLs were not consistent in their mode of action.
The minimum contig size considered for de novo assembly into scaffolds was bp. The N50 scaffold length for the Kiroba assembly was 3, bp Figure 4. N50 length is the length of the smallest contig scaffold for which the collection of all contigs of that length or longer contains at least half of the sum of the lengths of all contigs.
The final assembly had The Kiroba genome assembly was of sufficient quality for comparative analysis. Figure 4. Kiroba genome assembly showing the contig size in log The minimum size used in the scaffold assembly was bp blue line and the N50 3, bp of the contigs assembly red. Large M. The introgression region in chromosome 18 encodes many protein domains that are associated with defense, including kinases, F-box family protein which contain leucine-rich repeats LRR , tetratricopeptide repeat TPR -like superfamily protein, protein kinase superfamily protein, and pentatricopeptide repeat PPR superfamily protein.
Figure 5. Orange indicates M. The M. The results extend an earlier study Bredeson et al. Figure 6. Parent—offspring relationship is represented by a solid line, full-siblings by dotted lines and identicals by sinewave modified from Bredeson et al. Significant terms and their associated annotations related to disease resistance within the QTL regions were obtained. They are significant by P -value and false discovery rate FDR correction.
A total of significant terms were detected within the QTL regions and 17 6. The significant genes found on chromosome XI were not related to disease resistance in plants. There are only seven genes between the flanking markers of QTL on chromosome XVII, and none are particularly associated with disease resistance. The mapping population used for QTL analysis was developed from a cross between Kiroba as the female parent and AR as the male parent. Kiroba is a local landrace found in coastal Tanzania and shows strong field resistance toward CBSD in that it gets infected by the viruses but shows no or minimal root necrosis and mild foliar symptoms even after two years under high disease pressure.
Virus levels are kept low limiting the impact of the disease on yield Kaweesi et al. Breeding in cassava is technically challenging due to its heterozygous nature, long growing cycle and low seed yield per pollination. It is highly outcrossing and difficult to develop an adequate sized F 2 population usually limiting genetic studies to F 1 progenies Kunkeaw et al. This is likely to be due to the fact that seeds were germinated in pots on benches, exposing the soil to high diurnal temperature ranges.
Placing the seed trays on the ground, thereby reducing diurnal temperature ranges that the soil and seed were exposed to, increased germination rates results not shown. SSR analysis and identity by state analysis of SNP data revealed patterns of relatedness, unrelatedness and some uncertainties of some of the F 1 progeny. Some F 1 progeny were assigned to different populations, as they were probably off-types, most likely having received pollen from a source other than from the donor parent during crossing.
The number of progeny with unexpected alleles that were regarded as off-types was surprisingly high This drastically reduced the number of genotypes available for linkage and QTL mapping in this study. Recent observations indicate female flowers may remain receptive for some time even after pollination and introduction of pollen from elsewhere may have been possible by other pollinating agents such as bees. The present assumption of non-receptivity of the flowers after pollination needs to change and the crossing technique modified to include bagging of the flowers even after pollination.
CBSD disease distribution varied with site and season. CBSD foliar symptoms were higher in Chambezi for both phenotyping years; — and — mean score 2. CBSD root necrosis was also higher in Chambezi for the two years mean score 3. All classes of scores 1—5 were represented in Chambezi for both leaf and root symptoms, but only classes 1—3 in Naliendele. The presence of different strains of virus in Chambezi or an increased population of whitefly may have contributed to this observation Ndunguru et al.
The phenotypic data in this study generally do not follow a perfect normal distribution and this is not surprising as it is a common phenomenon observed in many mapping populations Zou, CBSD foliar symptoms were positively skewed toward class1 in all the years in Naliendele skewness 3.
This is supported by the kurtosis analysis showing Naliendele being positively skewed kurtosis CBSD root necrosis follows a similar trend; Naliendele positively skewed skewness, 2.
This shows the disease pressure and distribution was high in Chambezi in all the years but much lower in Naliendele. The genetic map was based on a one-step map approach Rabbi et al. Bridge markers are needed in each linkage group for the maps to be integrated Tang et al.
The number of times that individual genes have been identified following a QTL mapping study remains small. Indeed, Roff lists examples of quantitative traits in which single genes have major effects and their molecular basis has been studied, and he notes that this number is modest relative to the effort invested in QTL studies.
One reason for this discrepancy is that many QTL map to regions of the genome of perhaps 20 centimorgans cM in length, and these regions often contain multiple loci that influence the same trait see, however, Price, Moreover, identifying the actual loci that affect a quantitative trait involves demonstrating causality using techniques like positional cloning see Clee et al. Frequently, the quest for individual genes within a QTL is assisted by the identification of a priori candidate genes using classical reverse genetics or bioinformatics.
A functional relationship between the candidate gene and the QTL must then be demonstrated, such as by using functional complementation the addition of wild-type complementary DNA from the gene in question into the nucleus to rescue a loss-of function mutation or to produce an alternative phenotype; see, for example, Frary et al.
Other techniques, such as deficiency mapping deletion mapping , are available for specific organisms, including Drosophila Mackay, New permutations of QTL mapping build upon the utility of the original premise: locus discovery by co-segregation of traits with markers. Now, however, the definition of a trait can be broadened beyond whole-organism phenotypes to phenotypes such as the amount of RNA transcript from a particular gene expression or eQTL; Schadt et al.
QTL mapping works in these contexts because these phenotypes are polygenic , just like more traditional organismal phenotypes, such as yield in corn. For example, transcript abundance is controlled not just by cis -acting sequences like the promoter , but also by potentially unlinked, trans -acting transcription factors.
Similarly, protein abundance is controlled by "local" variation at the coding gene itself, and by "distant" variation mapping to other regions of the genome. Local variation is likely to be composed of cis variants controlling transcript levels though the correlation between transcript level and protein abundance is often quite low, so this may represent a minority of cases; see Foss et al. Other local mechanisms might include polymorphisms for the stability or regulation of the protein.
In contrast, distant variation could include upstream regulation control regions. Beyond these examples, further extension of QTL analysis includes mapping the contribution of imprinting to size-related traits Cheverud et al. Historically, the availability of adequately dense markers genotypes has been the limiting step for QTL analysis. However, high-throughput technologies and genomics have begun to overcome this barrier. Thus, the remaining limitations in QTL analysis are now predominantly at the level of phenotyping, although the use of genomic and proteomic data as phenotypes circumvents this challenge to some extent.
Genome-wide association studies GWAS are becoming increasingly popular in genetic research, and they are an excellent complement to QTL mapping. Whereas QTL contain many linked genes , which are then challenging to separate, GWAS produce many unlinked individual genes or even nucleotides, but these studies are riddled with large expected numbers of false positives.
Though GWAS remain limited to organisms with genomic resources, combining the two techniques can make the most of both approaches and help provide the ultimate deliverable: individual genes or even nucleotides that contribute to the phenotype of interest.
Indeed, combining different QTL techniques and technologies has great promise. For example, Hubner and colleagues used data on gene expression in fat and kidney tissue from two previously generated, recombinant rat strains to study hypertension. Alternatively, samples adapted to different environments may be compared, or other populations of interest might be selected for expression analysis.
This approach permits measurement of hundreds or even thousands of traits simultaneously. Other interesting questions concerning gene regulation can be addressed by combining eQTL and QTL, such as the relative contributions of cis -regulatory elements versus trans -regulatory elements.
Regarding hypertension, Hubner et al. These integrated approaches will become more common, and they promise a deeper understanding of the genetic basis of complex traits, including disease Hubner et al. Integrating phenotypic QTL with protein QTL can also give investigators a more direct link between genotype and phenotype via co-localization of candidate protein abundance with a phenotypic QTL De Vienne et al.
Still more kinds of data can be integrated with QTL mapping for a "total information" genomics approach e. QTL studies have a long and rich history and have played important roles in gene cloning and characterization; however, there is still a great deal of work to be done. Existing data on model organisms need to be expanded to the point at which meta-analysis is feasible in order to document robust trends regarding genetic architecture.
Furthermore, QTL studies can inform functional genomics , in which the goal is to characterize allelic variation and how it influences the fitness and function of whole organisms. Thus, although the map between genotype and phenotype remains difficult to read, QTL analysis and a variety of associated innovations will likely continue to provide key landmarks. Albert, A. The genetics of adaptive shape shift in stickleback: Pleiotropy and effect size.
Evolution 62 , 76—85 Baack, E. Selection on domestication traits and quantitative trait loci in crop-wild sunflower hybrids. Molecular Ecology 17 , — Beavis, W.
QTL analyses: Power, precision, and accuracy. In Molecular Dissection of Complex Traits , ed. Casa, A. Proceedings of the National Academy of Sciences 97 , — Cheverud, J. Genomic imprinting effects on adult body composition in mice. Proceedings of the National Academy of Sciences , — Clee, S. Positional cloning Sorcs1, a type 2 diabetes quantitative trait locus. Nature Genetics 38 , — link to article. Damerval, C. Quantitative trait loci underlying gene product variation—A novel perspective for analysing regulation of genome expression.
Genetics , — Darvasi, A. Experimental strategies for the genetic dissection of complex traits in animal models. Nature Genetics 18 , 19—24 link to article. De Vienne, D. Genetics of proteome variation for QTL characterization: Application to drought-stress responses in maize.
Journal of Experimental Botany 50 , — Doebley, J. Teosinte branched 1 and the origin of maize: Evidence for epistasis and the evolution of dominance. Falconer, D. Introduction to Quantitative Genetics , 4th ed. London, Prentice Hall, Forbes, S. Quantitative trait loci affecting life span in replicated populations of Drosophila melanogaster.
Composite interval mapping. Foss, E. Genetic basis of proteome variation in yeast. Nature Genetics 39 , — link to article. Frary, A. Science , 85—88 Gupta, P. Functional and Integrative Genomics 4 , — Hayes, B. The distribution of the effects of genes affecting quantitative traits in livestock. Genetics Selection Evolution 33 , — Hubner, N. Integrated transcriptional profiling and linkage analysis for identification of genes underlying disease. Nature Genetics 37 , — link to article.
Novel integrative approaches to the identification of candidate genes in hypertension. Hypertension 47 , 1—5 Kearsey, M. The principles of QTL analysis a minimal mathematics approach.
Journal of Experimental Botany 49 , — Lark, K. Meanwhile, with regard to linseed flax, China is the third largest source after India and Canada, with the United States coming in fourth. Optimal flax cultivars for textile production can yield flax of high quality, high yield, high fiber content, and other desirable characteristics. However, compared to other crops, little systematic in-depth research has been focused on improving flax cultivars for use with traditional breeding methods.
Marker-assisted selection MAS would enhance the effectiveness of existing breeding methods to improve flax stem fiber, but currently insufficient molecular resources are available for achieving that goal. Indeed, in spite of the economic importance of flax, until now breeding programs utilizing MAS resources have only been used to construct genetic linkage maps for various crops such as rice Orjuela et al.
Fortunately, MAS resources of other crops can provide a valuable foundation of knowledge for future molecular biological studies of flax. Until recently, flax improvement has relied mainly on conventional breeding methods based upon limited germplasm resources. However, most important agronomic characteristics, such as stem fiber content, are quantitative innate traits that are regulated by micro-effect polygenes and environmental factors.
Because increasing the fiber content is an ultimate goal in flax genetics and breeding research, identification of quantitative trait loci QTL would be advantageous for MAS and map-based cloning. Unfortunately, a linkage map with suitable marker density as a prerequisite for QTL detection does not yet exist Thoday, ; Paterson et al. Indeed, only four limited flax linkage maps and one integrated low-density map have been published Spielmeyer et al. Moreover, integration of these maps to generate a draft flax genome would require additional information before it could be used to effectively support MAS-based breeding.
Therefore, the goal of this study was to construct a suitably dense flax genetic linkage map composed of previously published flax genome assemblies Wang Z. These genetic resources should lay a foundation for further mining of flax genes important for optimal fiber generation for use in MAS-based breeding of flax.
A total of The GC guanine-cytosine content was For the analysis of the F 2 mapping population, 1,,—3,, reads were generated for the development of 67,—90, SLAF markers for each plant, with marker coverage ranging from 3. Figure 1. Coverage and number of markers for each of the F2 individuals. The x-axes in both A,B indicate each of the F 2 individuals; the y-axes indicate the marker coverage A and the number of markers developed for each F 2 plant B.
After correcting or discarding low-depth SLAF tags, , high-quality SLAFs were identified, of which 23, were polymorphic, for a polymorphism rate of The remaining , SLAFs were non-polymorphic or repetitive. Figure 2. Number of polymorphic SLAF markers for eight segregation patterns. The x-axis indicates eight segregation patterns of polymorphic SLAF markers; the y-axis indicates the number of markers.
After removing incomplete and significant segregation distortion markers, 7, SLAFs were retained for the final genetic linkage map construction. The mLOD values were calculated between pairs of tags following a standard procedure as previously described Vision et al.
HighMap software Liu et al. Subsequently, a genetic linkage map of overall length 1, The basic characteristics of all linkage groups LGs obtained are shown in Table 3. LG11 contained the maximum number of markers , while LG12 contained the minimum marker number The genetic length of the 15 linkage groups ranged from Figure 3. Distribution of SLAF markers on the 15 linkage groups of flax.
A black bar indicates a SLAF marker. The x-axis represents the linkage group number and the y-axis indicates the genetic distance cM within each linkage group. The quality of the flax genetic map was evaluated by analyzing the integrity of mapped markers. The average integrity of each individual marker was Figure 4. The integrity distribution map of all individuals. The x-axis represents the individuals and y-axis represents the complete degree of mapped markers.
Haplotype maps, which reflect chromosomal-exchange events among genomes within a population, were developed for the offspring and parental controls using the 2, SLAF markers Presentation S1. Dual-exchange sites might result from two scenarios: 1 a recombinant hotspot region within the genome and 2 genotyping error caused by sequencing within a linkage group. Consequently, a higher proportion of dual-exchange events correlated with greater success of genotyping and more effective ordering of markers.
In this study, the percentage of dual-exchanges ranged from 0. Ultimately, most recombination blocks were clearly defined and the LGs were uniformly distributed, suggesting that genetic mapping was of high quality. A genetic map essentially reflects multipoint recombination analysis, with closer distances between adjacent markers reflecting smaller observed recombination rates. To analyze recombination relationships between markers, we determined the potential layout of mapped markers.
The quality of the genetic map was also evaluated using heat maps which directly depicted recombination relationships among markers for all fifteen linkage groups Presentation S2. Each cell of the heat map represents a recombination rate between two adjacent markers whereby the rate level was depicted using different colors ranging from yellow to purple yellow indicating a lower recombination rate; purple indicating a higher rate.
In this way, heat maps were generated for each LG using recombination scores for pair-wise comparative analyses of the 2, markers Presentation S2. Phenotypic data for F 2 families are presented in Table 7 and the frequency distribution of all measured traits is shown in Figure 5.
Plant height, stem length, and stem yield all strictly followed normal distributions, with kurtosis and skewing values close to zero. Meanwhile, the frequency distributions of seed yield and fiber content exhibited approximately normal distributions, but with higher kurtoses although only the left peaks were characterized.
The results indicate that quantitative traits are controlled by multiple genes, with overall trait values of offspring biased approaching value of the pure-bred parents. These results therefore suggest a heterosis degradative phenomenon may be operating for some traits in flax. Figure 5. Frequency distribution of flax fiber related traits. A : plant height; B : stem length; C : seed yield; D : stem yield; E : fiber yield; F : fiber content. Based on the assembly information of flax genome, which chromosome-scale pseudomolecules were published You et al.
Interestingly, main effect QTLs for plant height and fiber yield were both detected to the same extent on chrLG1 which spanned the genetic distance from A similar result was observed for stem length, although the different genetic distance spanned, whereby the main effect QTLs were detected to the same scaffold within chrLG5.
All of these anchored QTLs should serve as a foundation for later accurate identification of related genes.
Previously, flax QTL mapping based on the genetic linkage map alone could only successfully map a few markers or segments linked to a particular target trait gene, but could not localize target trait genes to physical chromosomal regions.
In this study using genomic assembly information in addition to QTL mapping, target trait QTLs could be identified on chromosomes and could be localized to smaller chromosomal segments.
However, because each segment may contain several or even dozens of genes, further localization and identification of target genes will still require great effort. Nevertheless, the results presented here provide a foundation for further flax gene mining and characterization.
Large-scale genotyping methodologies play an important role in genetic association studies. One such technology, SLAF-seq, has been applied in many plant studies and has produced remarkable results.
Due to its relatively higher density, excellent consistency, effectiveness, and lower cost when compared with traditional methods, this method has become a popular genotyping method. Subsequently, in this work sequencing results provided a large quantity of SLAF markers to further drive flax genomic research. Moreover, the SLAFs developed here may also be valuable in studies of other flax cultivars and offspring for identification of germplasm or hybrids and for analysis of genetic diversity among cultivars.
SLAF-seq has been widely successful due to several distinguishing characteristics: i deep sequencing ensures genotyping accuracy; ii reduced representation strategy reduces sequencing costs; iii a pre-designed reduced representation scheme optimizes marker efficiency; and iv a double barcode system facilitates sequencing of large populations Sun et al.
Consequently, this technology has been used in numerous genetic linkage map construction studies Zhang et al. In the present study, the total length of the new genetic linkage map, which was constructed using 2, SLAF markers, was 1, To our knowledge, this map is the highest density genetic map of flax currently available and is of high quality. Although sequencing completed to date toward obtaining the entire flax genome sequence has produced high-quality genomic information, assembly has been difficult due to the high complexity of the genome.
However, much work is still needed before the entire flax genome can be assembled, due to genomic complexity and a large number of repeated regions within the genome Wang Z. As the carrier of flax genetic information, the genome is the basis for research on genetic mechanisms of flax.
Fortunately, chromosome-scale pseudomolecules were refined by optical, physical, and genetic maps in flax You et al. We employed the genome assembly information, with the high-density map of the flax genome constructed here, 12 QTLs were detected and linked with the chromosome scaffolds.
Although the order is different in chrLGs and chromosome-scale pseudomolecules, it still in conjunction with sequence information to generate 15 chromosomes. Notably, this method has generated ideas and strategies with wider applicability for use in whole genome assembly of related species.
Interestingly, characteristic fiber-related traits were observed to segregate in flax F 2 offspring. Moreover, because plant phenotypic traits of height, stem length, stem yield, seed yield, fiber yield, and fiber content exhibit characteristics of quantitative traits, the mapped QTLs of these fiber-related traits to physical locations on chromosome scaffolds.
Moreover, the results have helped us make provisional inferences of pleiotropic genes or neighboring genes that influence plant height and fiber yield to explain observed morphological correlations.
0コメント