Estimation of parental genome contributions among doubled haploid maize lines derived from six backcross populations.

Abstract

Doubled haploid (DH) lines have increasingly been used in maize research and breeding due to substantial progress in the in vivo haploid induction technology and the availability of a dominant anthocyanin marker gene for haploid identification. Although DH lines are supposed to be genetically pure, preliminary field observation indicated some level of non-uniform plant characteristics within seed lot of some DH lines. Molecular characterization of the DH lines and estimating their parental genome contributions would be useful for germplasm organization and choosing pairs of DH lines for hybrid make up or new pedigree starts. Next generation genotyping by sequencing (GBS) provides a low cost and high marker density platform for such purposes. The objectives of the present study were, therefore, to i) understand the genetic purity, genetic distance and relationship among 417 maize DH lines derived from 6 BC, populations using GBS; ii) estimate the parental genome contributions to the genome of the DH lines derived from different genetic backgrounds; and iii) check the correlation between parental genome contributions and grain yield under drought and optimum environments. A total of 97190 markers that were polymorphic across the 417 DH lines were used for the genetic purity, genetic distance, and relationship analyses. The number of polymorphic markers per chromosome varied from 5752 on chromosome 9 to 13643 on chromosome 1, with a mean of 9719. A smaller subset of 29074 markers, ranging from 3724 to 6665 per population, was used for estimating parental genome contributions. About 97% of the DH lines (404 out of 417) were genetically pure, with <2% heterogeneity (the number of markers that were not homozygous due to admixture or contamination). Only 2 DH lines showed heterogeneity >5%, which require further purification. Genetic distance between pairwise comparisons of the 417 DH lines ranged from 0.055 to 0.457 and only 13 out of 86,736 pairwise comparisons (0.02%) showed a distance <0.100, which clearly suggests the presence of large genetic differences among the DH lines. Both populations 1 and 6 showed significantly lower (p<0.001) donor introgression than the other four populations. Donor parent contribution was significantly(p <0.001) higher in the CML444 genetic background than both CML395 and CML488.The average donor and recurrent parent genome contributions across the 417 DH lines was 31.7% and 64.3%, respectively. About 82% of the DH lines showed donor genome introgression higher than the expected 25% at BC, generation, which may be due to artificial selection during the DH process (induction, genome doubling, and seed increase), during the development of F, or BC, seeds, or during initial agronomic evaluation of the DH lines. Donor parent introgression up to 32% showed significant positive correlation with grain yield under drought (r=0.312, p <0.001) and optimum (r = 0.142, P <0.050) environments. Such correlation was not observed when donor parent introgression exceeds 32%, which may be due to simultaneous transfer of undesirable genes flanking the genes of interest from the donor parents. As the DH-derived hybrids were evaluated for phenotypic performance only at one managed drought screening environment, additional data may be needed to confirm the correlations reported in this study and also map the specific quantitative trait loci (QTL) associated with drought tolerance and hence grain yield.