This dataset includes marker effects from four maize biparental crosses (A/B, C/B, C/D, and E/F) and for five traits: grain yield (t ha⁻¹ at 150 g H₂O kg⁻¹), grain moisture (g kg⁻¹), test weight (g L⁻¹), plant height (cm), and ear height (cm). Marker effects were estimated separately for two testers per population. Testers are designated as T1, T2, T3, and T4.

Four maize biparental crosses (A/B, C/B, C/D, E/F) with 130 to 242 doubled haploid (DH) lines were crossed to two inbred testers, which were designated as T1, T2, T3, and T4. The eight testcross populations were evaluated in up to 18 year-location combinations in the U.S. Corn Belt. The measured traits included grain yield (t ha^-1 at 150 g H₂O kg^-1), grain moisture (g kg^-1), test weight (g L^-1), plant height (cm), and ear height (cm). Testcross phenotypic data were adjusted according to the incomplete block design used at each location to obtain best linear unbiased estimates (BLUEs) per location. Across-environment BLUEs were subsequently calculated. Variance components were estimated with testcross and environment effects treated as random.

The DNA marker data consisted of 3072 single nucleotide polymorphisms (SNPs) generated using the Illumina iScan platform (Illumina Inc., San Diego, CA). Across all population, SNPs with > 10 % missing values, > 5 % of heterozygosity, or minor allele frequency < 0.1 were excluded. Missing marker data were imputed within each population by using the most correlated SNP and declaring SNP loci with allele frequencies outside the expected range for a DH population (<0.25 or >0.75) as monomorphic for the major allele. Marker effects were obtained separately for the polymorphic markers of each biparental cross by ridge regression-best linear unbiased prediction (RR-BLUP) via a modified version of the RRBLUP software (Bernardo 2020).