The Strategy and Potential Utilization of Temperate Germplasm for Tropical Germplasm Improvement—a Case Study In Maize (Zea mays. L).

Organization of maize germplasm into genetically divergent heterotic groups is needed for breeding high yielding hybrids by exploitation of heterosis. In this study, 94 CIMMYT inbred lines (CML) and 54 US Germplasm Enhancement of Maize (GEM) lines were assembled and characterized using 1,266 SNPs with high quality. A total of 321 Tuxpeño core subset (618 individual plants) were genotyped using 1,041 out of the 1,266 SNPs. Based on principle component analysis (PCA) using 1,041 SNPs, GEM, CML, and Tuxpeño germplasms were clearly separated. There were two groups of GEM lines classified by PCA that seemed to correspond to the stiff stalk (SS) and non-stiff stalk (NSS) heterotic groups bred by GEM. CML lines did not form obvious subgroups by PCA.  Each of 1,266 SNP allele frequency differed in GEM and CML. 3.6% alleles (46/1,266) of CML were absent in GEM, and 4.4% alleles (56/1,266) of GEM were absent in CML. 13 and 35 alleles existing in Tuxpeño germplasm were absent in GEM and CML, respectively. CML heterotic group A with GEM heterotic group SS and CML heterotic group B with GEM heterotic group NSS were crossed and developed 654 F1. Association mapping was performed based on both 148 CML, GEM lines and 654 F1s for seven agronomic traits at two locations. For plant height (PH) and anthesis time (AT), genomic estimated breeding values (GEBVs) for a testing set of 215 F1s were predicted based on the training data of 430 F1s using a best linear unbiased prediction (BLUP) method. The accuracy benefitted from the adoption of markers associated with QTLs for both traits. However, it does not necessarily increase with the raise of associated marker density. High correlations between actual and predicted phenotypic values of F1 indicated the possibility to choose high performing parental combinations of phenotypic traits using SNP markers. It is hoped that we can enhance the tropical maize germplasm by incorporating temperate germplasm more efficiently, through genome wide selection. Clear heterotic patterns of the GEM lines and unique alleles, in addition to useful diversity in Tuxpeño core subset, can be exploited in tropical maize hybrid breeding.