UFYMA   27844
UNIDAD DE FITOPATOLOGIA Y MODELIZACION AGRICOLA
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Genome‑wide association and genomic prediction of resistance to Mal de Rio Cuarto disease from exotic germplasm and local phenotypic assessments
Autor/es:
BONAMICO NATALIA C.; ROSSI, EZEQUIEL A.; BALZARINI, MÓNICA; RUIZ, MARCOS
Lugar:
Córdoba
Reunión:
Simposio; Primer Simposio de Mejoramiento Genético Vegetal; 2021
Resumen:
Resistance to Mal de Rio Cuarto (MRC) disease is a quantitative trait. Traditional linkage mapping and association mapping have been successfully applied to identify genomic regions conferring resistance to MRC disease in maize. Genomic selection (GS) is a promising breeding tool. The major advantage of GS is that all polymorphisms affecting a trait are modeled, regardless of effect size, making it a potentially powerful approach for a complex trait like MRC resistance. The objectives of the study were (1) to identify genomic regions, SNPs associated with MRC resistance; and (2) to assess the potential of genomic selection (GS) for MRC resistance in maize.A set of 187 CIMMYT maize lines was phenotypical assessed for their response to MRC disease under natural infection. The trials were performed in five environments (locations-years combinations) belonging to the area where MRC is endemic in Argentina. A partially replicated (P-rep) design involved the use of three replicates for 25% of lines and single plots of the remaining 75% of lines, was used in all trials. A disease severity index (DSI) was calculated for each plot. The BLUPs of genotype effects across environments were used as phenotypes in the GWAS and GS analysis as response variable. From the public GBS SNP marker data, a sub-set of 10810 SNPs, with low missing data rate (0.25 were used for GWAS. The GWAS was carried out by GAPIT package in R software. The genomic prediction was carried out by ridge regression best linear unbiased prediction model (RR-BLUP). Two approaches were performed, a model without inclusion of GWAS based MRC resistance associated SNPs and another model with inclusion of associated SNPs. Prediction accuracy of the GS approach was evaluated using the five-fold cross-validation with 100 times repetitions. The prediction accuracy was calculated by dividing the correlation between the predicted genotypic values and observed phenotypic values by the square root of heritability. GWAS results allowed us to identify 12 significant markers-trait associations for MRC resistance (P < 1 × 10-3). These significantly associated SNPs individually explained 6-11% of the total genotypic variance. The prediction accuracy was 0.38 in the model without MRC associated SNPs and 0.39 with the inclusion of MRC resistance associated SNPs into the prediction model (Fig. 2). Differences in accuracies between model without associated SNPs and model with the inclusion of associated SNPs were not large, indicating that the polygenic small effects are important in the genetic architecture underlying the resistance to MRC. The prediction accuracy of the genomic prediction is acceptable given the complexity of the MRC resistance. The inclusion of MRC associated SNPs into prediction model led only slight increase in the prediction accuracy, indicating that prediction is mainly attributable to many small effects genomic regions distributed across genome. The results from this study give first insights into the potential of genomic prediction of MRC resistance in maize.