Search

Haplotype blocks might carry additional information compared to single SNPs and have therefore been suggested for use as independent variables in genomic prediction. Studies in different species resulted in more accurate predictions than with single SNPs in some traits but not in others. In addition, it remains unclear how the blocks should be built to obtain the greatest prediction accuracies. Our objective was to compare the results of genomic prediction with different types of haplotype blocks to prediction with single SNPs in 11 traits in winter wheat. We built haplotype blocks from marker data from 361 winter wheat lines based on linkage disequilibrium, fixed SNP numbers, fixed lengths in cM and with the R package HaploBlocker. We used these blocks together with data from singleyear field trials in a cross-validation study for predictions with RR-BLUP, an alternative method (RMLA) that allows for heterogeneous marker variances, and GBLUP performed with the software GVCHAP. The greatest prediction accuracies for resistance scores for B. graminis, P. triticina, and F. graminearum were obtained with LD-based haplotype blocks while blocks with fixed marker numbers and fixed lengths in cM resulted in the greatest prediction accuracies for plant height. Prediction accuracies of haplotype blocks built with HaploBlocker were greater than those of the other methods for protein concentration and resistances scores for S. tritici, B. graminis, and P. striiformis. We hypothesize that the trait-dependence is caused by properties of the haplotype blocks that have overlapping and contrasting effects on the prediction accuracy. While they might be able to capture local epistatic effects and to detect ancestral relationships better than single SNPs, prediction accuracy might be reduced by unfavorable characteristics of the design matrices in the models that are due to their multi-allelic nature.

Summary Genome structural variation (SV) contributes strongly to trait variation in eukaryotic species and may have an even higher functional significance than single‐nucleotide polymorphism (SNP). In recent years, there have been a number of studies associating large chromosomal scale SV ranging from hundreds of kilobases all the way up to a few megabases to key agronomic traits in plant genomes. However, there have been little or no efforts towards cataloguing small‐ (30–10 000 bp) to mid‐scale (10 000–30 000 bp) SV and their impact on evolution and adaptation‐related traits in plants. This might be attributed to complex and highly duplicated nature of plant genomes, which makes them difficult to assess using high‐throughput genome screening methods. Here, we describe how long‐read sequencing technologies can overcome this problem, revealing a surprisingly high level of widespread, small‐ to mid‐scale SV in a major allopolyploid crop species, Brassica napus. We found that up to 10% of all genes were affected by small‐ to mid‐scale SV events. Nearly half of these SV events ranged between 100 bp and 1000 bp, which makes them challenging to detect using short‐read Illumina sequencing. Examples demonstrating the contribution of such SV towards eco‐geographical adaptation and disease resistance in oilseed rape suggest that revisiting complex plant genomes using medium‐coverage long‐read sequencing might reveal unexpected levels of functional gene variation, with major implications for trait regulation and crop improvement.

Additional file 3. Review history.

Additional file 2. Single plant selection shopping list.

Additional file 1: Supplementary Figure 1. Scatter plot representing DA loading of Ugandan accessions based on races. (a) and (b) are replicate 2 and 3 respectively. Each dot represents an individual and the colour code is displayed in the index.

Since the discovery of the tobacco mosaic virus in the 1890s, awareness has grown in regard to how viruses affect the environment. Viral infections are now known to cause various effects besides pathogenicity, with some viruses in fact having a beneficial impact on plants. Although research has focused on disease-causing viruses that can infect plants, many wild plants are also infected with non-pathogenic viral agents. Traditionally, abiotic, and biotic stresses have been studied as isolated stimuli that trigger signaling pathways within the plant. However, both biotic and abiotic stress can trigger complex molecular interactions within plants, which in turn drive interconnected response pathways. Here, we demonstrate that heat-killed tobacco mosaic virus (TMV) can increase abiotic stress tolerance in plants, an effect that could potentially be implemented in challenging growth environments. To our knowledge, this is the first report of plant abiotic stress tolerance following treatment with heat-killed viral particles.

Additional file 1: Fig. S1. Overview of experimental and research analysis methods. Fig. S2. Venn diagram of the distribution of genes regulated by different types of eQTLs (local eQTL and distant eQTL) at 20 DAF and 40 DAF. (a) Distribution number of genes which were regulated by different types of eQTLs at 20 DAF. (b) Distribution number of genes which were regulated by different types of eQTLs at 40 DAF. Fig. S3. Manhattan plot of BnaA05.FAD7 eGWAS at 40 DAF. Fig. S4. Study design on ATAC-seq of 6 representative accessions of B. napus. Fig. S5. Correlation analysis of 59 ATAC-seq samples. The samples are named according to “(22, 26, 34 or 40) DAF” + “accession_cellular ploidy (2C, 3C or 4C)” + “biological replicate” format naming. Fig. S6. Regional plot of ATAC-seq data and eGWAS results of BnaA08.TGD1. BnaA08.TGD1 is marked by a dashed line. The shaded area indicates the lead SNP of the local eQTL affecting BnaA08.TGD1. Fig. S7. Comparison of the explained variance (r2) of eQTLs for expression variation in or not in OCRs. Fig. S8. Expression correlation analysis of adjacent genes and randomly sampled gene pairs. Violin plot shows that the expression correlation of adjacent genes is significantly higher than that of randomly sampled gene pairs, *** indicates P < 0.001 in Kolmogorov-Smirnov test. Fig. S9. eQTL localization of B. napus. (a) Dot plot showing eQTL and their regulated genes in 19 chromosomes. x-axis indicates the physical position of each variant on the ZS11 genome. yaxis indicates the physical position of the localized gene on the ZS11 genome and each point indicates a detected eQTL locus. Points on the diagonal line indicate local eQTLs and points away from the diagonal line indicate distant eQTLs. (b) Average number of eQTL for different gene types. Fig. S10. The propensity of genome-wide homoeologous gene expression on different chromosomes of An and Cn at 40 DAF. Fig. S11. Effect of local eQTLs on asymmetric regulation (Cn gene regulating An gene) of subgenomes. Fig. S12. Comparison of gene correlations in different types of aHGPs at 40 DAF. “BG” represents all HGPs. “Local” represents aHGPs with local eQTLs and “No Local” represents aHGPs without local eQTLs. Fig. S13. The density map shows the enrichment of TFs in HGPs with feedback regulation. The blue inverted triangle represents 333 HGPs that are TFs. Fig. S14. The density plot shows the maximum number of eQTLs within 1 Mb. We defined the number of eQTLs within 1 Mb greater than 81 (P < 0.01) as a hotspot (red dot). Fig. S15. Characterization of SOC-related gene regulation in subgenomic imbalance. a Comparison of SOC-related gene correlations in different types of aHGPs at 20 DAF. “BG” represents all HGPs. “Local” represents aHGPs with local eQTLs and “No Local” represents aHGPs without local eQTLs. b Correlation between expression levels of BnaA02.PDAT1 and BnaC02.PDAT1 at 20 DAF. c Correlation between expression levels of BnaA01.AAD5 and BnaC01.AAD5 at 20 DAF. d Correlation between expression levels of BnaA07.WRI1 and BnaC06.WRI1 at 20 DAF. Fig. S16. The enrichment of hotspots regulating TWAS significant genes of SOC. The x-axis and y-axis represent enrichment of TWAS significant genes of SOC in hotspots at 20 DAF and 40 DAF. Fig. S17. Workflow of XGBoost module. The collected expression data of TFs were used to construct the XGBoost model, upstream TFs prediction is performed for each gene in a gene set, and finally the prediction results of the whole gene set are summarized. Fig. S18. Expression of BnaA09.SCL31 at 20 DAF is regulated by a distant hotspot eQTL, which in turn regulates downstream SOC-related genes and affects seed oil content. Box plots of SOC and SOC-related genes (BnaA07.NAC13, lysophosphatidic acid acyltransferase (BnaC07.LPAAT), GLYCOLIPID TRANSFER PROTEIN (BnaC06.GLPT), SHEWENELLA-LIKE PROTEIN PHOSPHATASE 1 (BnaC03.SLP1)) based on haplotypes of distant eQTL. Fig. S19. The population information of BnaA09.SCL31 expression level at 20 DAF. a Box plots for expression levels at 20 DAF based on the haplotypes of variants in the gene region and the upstream 2 kb region of BnaA09.SCL31. b Correlation between SOC and expression levels of BnaA09.SCL31 at 20 DAF. Fig. S20. Identification of Arabidopsis T-DNA Mutants of SCL31. (a) Schematic diagram of TDNA insertion sites in SCL31 mutants (scl31-1, scl31-2). (b) Identification of homozygous TDNA mutants by PCR with a pair of gene specific primers (LP+RP) or combination of T-DNA border primer (LB1.3) and gene specific primers (RP). (c) Expression of SCL31 in WT, scl31-1, scl31-2 analyzed by quantitative RT-PCR using RNA extracted from leaves. Values are means(SD) (n = 3 biological repeats). ** indicates P < 0.01 and *** indicates P < 0.001 compared with WT in Student’s t test. Fig. S21. Fatty acid composition of SCL31 mutant seeds (scl31-1, scl31-2). Values are means(SD) (n = 5 biological repeats). * indicates P < 0.05 and ** indicates P < 0.01 compared with WT in Student’s t test. Fig. S22. Silique and seed of SCL31 mutant lines (scl31-1, scl31-2). (a) Silique of scl31-1 and scl31-2. Bar = 2mm. (b) Seed of scl31-1 and scl31-2. Bar = 500𝜇m. (c) Seed length of scl31-1 and scl31-2. Values are means(SD) (n = 5 biological repeats) and different letters indicate differences at P < 0.05 using Student’s t test. (d) Seed width of scl31-1 and scl31-2. Values are means(SD) (n = 5 biological repeats) and different letters indicate differences at P < 0.05 using Student’s t test. Fig. S23. Workflow of Basenji module. ATAC-seq data are used to identify OCRs and construct basenji models. Subsequently, 2 kb promoter sequences of genes in the gene set are collected for key motif identification. Fig. S24. The performance metrics of the Basenji deep learning model trained with the ZS11 reference genome and 59 rapeseed ATAC-seq samples. The horizontal axis represents 59 different samples and the vertical axis represents the performance data of the corresponding fold (where r2 represents the square of the correlation. AUPRC is the accuracy between the predicted regression value and the true value using the precision score function in scikit-learn package to calculate the accuracy between the predicted regression value and the true value; AUROC is the area under the precision-recall curve of the predicted regression value and the true value using the roc_auc_score function in the in scikit-learn package). Fig. S25. The correlation between BnaA07.NAC13 and the expression of the downstream genes potentially regulated by BnaA07.NAC13. Fig. S26. Transcriptional regulation of BnaC04.MYC70 and BnaC02G0181500ZS are activated by BnaA07.NAC13. a Schematic representation of the constructs used for the dual-luciferase assay. The effector constructs contain BnaA07.NAC13 driven by the CaMV35S promoter. The reporter construct contains the firefly luciferase driven by BnaC04.MYC70 and BnaC02G0181500ZS promoter, and the Renilla luciferase (REN) driven by the CaMV35S promoter. And the black square is the terminator. b,c Bar graph showing the relative LUC/REN ratio in the dual-luciferase assay. Values are means(SD) (n = 3 biological repeats). Fig. S27. Expression of BnaA07.NAC13 and BnaC07.NAC13 in WT and OE lines analyzed by quantitative RT-PCR using RNA extracted from leaves. (a, b) Values are means(SD) (n = 3 biological repeats). *** indicates P < 0.001 compared with WT in Student’s t test. Fig. S28. Fatty acid composition phenotype of OE lines (BnaA07.NAC13, BnaC07.NAC13). (a, b) Values are means(SD) (n = 5 biological repeats). * indicates P < 0.05, ** indicates P < 0.01 and *** indicates P < 0.001 compared with WT in Student’s t test. Fig. S29. Manhattan plot of BnaA07.NAC13 eGWAS at 20 DAF and 40 DAF. Fig. S30. Correlation between expression levels of BnaA07.NAC13 and TFs (such as BnaA09.SCL31 , WIP DOMAIN PROTEIN 3 (BnaA09.WIP3), CRY2-INTERACTING BHLH 4 (BnaA09.CIB4), BLUEJAY (BnaA09.BLJ)) in Hotspot87-88 at 20 DAF and 40 DAF. Fig. S31. Correlation between expression levels of BnaA09.SCL31 and BnaA07.NAC13 at 20 DAF and 40 DAF. Fig. S32. Size and number of LDs expressing BnaA09.SCL31 and BnaA07.NAC13s in tobacco leaves. a LD count expressing BnaA09.SCL31 and BnaA07.NAC13s by size (average diameter). Values are means(SD) (n = 3 biological repeats). Different letters indicate significant difference at P < 0.05, as determined by one-way ANOVA with Tukey’s post-test. b Quantification of LD sizes in leaf mesophyll cells. Circles represent size of LDs. Fig. S33. Expression of NAC13, SUD1, LPAAT, SCL30, SLP1 in WT, scl31-1, scl31-2 analyzed by quantitative RT-PCR using RNA extracted from leaves. (a-e) Values are means(SD) (n = 3 biological repeats). “NS” indicates P > 0.05, * indicates P < 0.05, ** indicates P < 0.01 and *** indicates P < 0.001 compared with WT in Student’s t test. Fig. S34. Transcriptional regulation of genes are activated by BnaA09.SCL31 or BnaA07.NAC13. a Schematic representation of the constructs used for the dual-luciferase assay. The effector constructs contain BnaA09.SCL31 and BnaA07.NAC13 driven by the CaMV35S promoter, respectively. The reporter construct contains the firefly luciferase driven by BnaA07.NAC13, BnaC07.LPAAT, BnaC03.SLP1 and BnaA07.SRO3 promoter, and the Renilla luciferase (REN) driven by the CaMV35S promoter. And the black square is the terminator. b Bar graph showing the relative LUC/REN ratio in the dual-luciferase assay. Values are means(SD) (n = 3 biological repeats). BnaA07.SRO3 (SIMILAR TO RCD ONE 3).

Additional file 2: Supplementary Figure 2. Topographic map of Uganda showing the elevation.

Additional file 1: Fig. S1. Exemplification of how the targeted root traits (seminal root angle and root biomass) were combined to develop wheat lines with different types of root systems (see Fig. 2b). Fig. S2. Backcrossing scheme for the development of elite wheat introgression lines combining seminal root angle and root biomass in different configurations. Purple boxes indicate the generations that were subjected to bi-directional selection for root traits using the SPS approach. The resulting BC2F4 lines were phenotyped for above-ground traits (plant height and flowering time) in the field. The BC2F4:5 lines were also characterised for seminal root traits in a replicated phenotyping experiment under controlled conditions to confirm differences in root traits compared to the respective recurrent parent. Fig. S3. Boxplots displaying (a) visual score and (b) root dry biomass for each genotype screened. Table S1. Summary of relevant traits of all the lines used to determine the feasibility of screening individual wheat plants for both seminal root angle and root biomass. Values are the mean of 6 replicates. QTL status: --/- = line carries the haplotypes associated with low root biomass (i.e. h1, h2 or h8 for haploblock b and h1 for haploblock a), whereas ++/+ = line carries both desirable haplotypes for high root biomass (i.e. h3 for haploblock b and h2 for haploblock a) [27, 28]. Table S2. Comparison between root:shoot (R:S) ratio means using the Fisher-LSD test for the panel of seventeen lines in Fig. 1d used for visual score for estimating root biomass non-destructively. The table displays significant differences between genotypes in R:S ratio (* Significant at the P ��� 0.05; ** Significant at the P ��� 0.01; ***Significant at the P ��� 0.001). In this panel were included the recurrent parents Borlaug100 and the three donors (SW107, SW309 and SW388) used to develop the introgression lines with different root configurations. Table S3. Details of the selected Borlaug100 introgression lines, including generation (FGen), pedigree, field above-ground measurements (plant height, days to flowering) and controlled environment root phenotypes (seminal root angle and root dry biomass).

Blackleg is one of the major fungal diseases in oilseed rape/canola worldwide. Most commercial cultivars carry R gene-mediated qualitative resistances that confer a high level of race-specific protection against Leptosphaeria maculans, the causal fungus of blackleg disease. However, monogenic resistances of this kind can potentially be rapidly overcome by mutations in the pathogen’s avirulence genes. To counteract pathogen adaptation in this evolutionary arms race, there is a tremendous demand for quantitative background resistance to enhance durability and efficacy of blackleg resistance in oilseed rape. In this study, we characterized genomic regions contributing to quantitative L. maculans resistance by genome-wide association studies in a multiparental mapping population derived from six parental elite varieties exhibiting quantitative resistance, which were all crossed to one common susceptible parental elite variety. Resistance was screened using a fungal isolate with no corresponding avirulence (AvrLm) to major R genes present in the parents of the mapping population. Genome-wide association studies revealed eight significantly associated quantitative trait loci (QTL) on chromosomes A07 and A09, with small effects explaining 3–6% of the phenotypic variance. Unexpectedly, the qualitative blackleg resistance gene Rlm9 was found to be located within a resistance-associated haploblock on chromosome A07. Furthermore, long-range sequence data spanning this haploblock revealed high levels of single-nucleotide and structural variants within the Rlm9 coding sequence among the parents of the mapping population. The results suggest that novel variants of Rlm9 could play a previously unknown role in expression of quantitative disease resistance in oilseed rape.