The availability of genome high density (HD) data provides new opportunities for combined population genomic prediction. An imputation from Nellore animals has been implemented to analyze, the impact of imputed data on heritability estimation (h2) and breeding values for residual feed intake (RFI), by singlestep GBLUP. Data sets of 607 Nellore bulls were used to impute a genotype bank composed of 1,466 Nellore animals, resulting in a dataset of 735,964 SNPs for all animals. The BLUPF90 package was used to perform quality control, resulting in 471,301 SNPs. Analysis to obtain h2 and breeding values for RFI were performed, using an animal model considering the single-step GBLUP. The BLUPF90 was efficient in processing the imputed file. The higher h2 of the RFI suggests an influence of the imputed data. The correlation of the breeding values indicates that the use of HD and 56K data result in similar breeding values.