Reliable methods for predicting heterosis will improve the efficiency of crossbreeding schemes by reducing their dependency on expensive field-tests of multiple pure-line combinations. When heterosis is due to dominance, heterosis is proportional to the squared difference in allele frequency (SDAF) between parental pure-lines (not necessarily homozygous). Hence, a linear model including regression on SDAF partitions crossbred phenotypes into pure-line values and heterosis, even without pure-line phenotypes. We regressed the phenotypes of crossbreds on the SDAF between parental lines. We used 53,421 SNP genotypes of 3,427 White Leghorn sires, allele frequencies of 6 White Leghorn dam-lines and cage-based records on egg number and egg weight of ~210,000 crossbred hens. SDAF predicted heterosis for egg number and weight with an accuracy of 0.5. Heterosis predictions allowed preselection of pure lines before field-testing, saving 50% of field-testing cost with only 4% loss in heterosis. The concept of using SDAF to predict heterosis can be extended to predict heterosis for individual sires. This will quantify genetic variation between sires from the same pure-line, and can increase heterosis in crossbred populations. We derived the theoretical expectation for heterosis due to dominance in crossbred offspring of individual sires: “heterozygousity excess.” Next, we predicted heterosis by regressing offspring performance on the heterozygosity excess. Between-line differences accounted for 99.0% of the total variance in predicted heterosis, while within-line differences among sires accounted for 0.7%. Despite the limited effect on these traits, the individual sire differences for expected heterosis could provide an additional tool for optimizing breeding strategies in crossbreeding programs. Finally, we ran single-SNP GWAS for crossbred hens in 3 ways: (1) across-line GWAS with phenotypic data on all crosses that were produced by the 4 sire-lines; (2) line-level GWAS for the crosses produced by each of the 4 sire-lines; and (3) within-cross GWAS for the sire-lines that showed significant SNPs. Heterosis in crossbred populations can be predicted from genotype data on the parental lines. Selecting SNPs that have dominance effects may improve this prediction.