Experimental designs that exploit family information can provide substantial predictive power in quantitative trait variant discovery projects. The a posteriori granddaughter design was applied to the US Holstein dairy cattle population. Twenty-nine trait-by-chromosomal segment effects were found with probabilities < 10⁻²⁰ that a segregating quantitative variant was detected by chance. Polymorphism genotypes for 79 grandsires and 16,236 sons were determined by imputation for 3,148,506 polymorphisms across the entire genome; 444 Holstein bulls had complete genome sequence, including 38 of the grandsires. Concordance between quantitative trait locus genotype and polymorphism was determined for all 29 effects. Complete concordance was obtained only for daughter pregnancy rate on chromosome 18 and protein percentage on chromosome 20. For each quantitative trait locus, effects of the 20 polymorphisms with the highest concordance scores for the analyzed trait were computed by stepwise regression. The effects for stature on chromosome 7, daughter pregnancy rate on chromosome 18, and protein percentage on chromosome 20 met the following 3 criteria: complete or nearly complete concordance, significance of the polymorphism effect after correction for all other polymorphisms, and a marker coefficient of determination that was > 50% of the total multiple-regression coefficients of determination for the 20 polymorphisms with highest concordance. An intronic variant SNP on chromosome 5 at position 93,945,738 explained 7% of the variance for fat percentage and 85% of the total variance explained by the multiple-marker regression. Variants identified in this study are likely to provide improved predictive power for genomic evaluation of dairy cattle.