CORESTA Meeting, Agronomy/Phytopathology, 2017, Santa Cruz do Sul, AP 27

Insights of whole genomic studies in Burley recurrent selection program

PADUA J.M.V.; PULCINELLI C.E.; FERREIRA R.A.C.D.; BOARETTO L.F.; BARBOSA P.K.A.; WEISS V.A.
Souza Cruz Ltda, British American Tobbaco, Group R&D, Rio Negro, Brazil

The whole genomic prediction (WGP) and Genome Wide Association Study (GWAS) is gaining every day more space in breeding programs in different species. To use the WGP, a “big” number of molecular markers covering the genome of the specie is required and by the statistical models it is possible to associate them with the phenotype in order to increase the efficiency of the program. In tobacco breeding these methodologies are not applied like in corn and soybean. So there are many questions about the use and the efficiency of these methods. In this way the objective of this study is to evaluate the possibility of the usage of WGP and GWAS in tobacco recurrent breeding programs. For this 193 S0:1 progeny from the Cycle “0” of Burley recurrent selection program were evaluated in lattice design with three replications in two environments. The characteristics evaluated were yield (kg of leaves per plot) and total alkaloid content. DNA samples were collected from all the progenies to obtain genotyping-by-sequencing (GBS) data. The Genomic Best Linear Unbiased Prediction (GBLUP) additive model was used with the cross validation to access the prediction accuracy (phenotype × genotype) of the model. After the data imputation and filter process, 3.185 markers were used to run the model. The prediction accuracy in the cross validation using 70 % of the genotypes to fit the model were below 20 %. The GWAS study showed peaks around reported genes in the literature. These preliminary results shows that to incorporate the WGP and GWAS in the breeding program more studies should be done.