QTL locating and genomic selection for tassel-related traits using F2:3 lineage haploids.

Tassel size affects the nutrient allocation synthesized by photosynthesis in maize, which in turn affects ears development and the yield components, such as kernel rows, kernel number per row, seeding rate, and kernels weight per one hundred. In this study, haploids were induced using the F2:3 families, which were constructed from the elite inbred lines Zheng 58 and B73. Genotypes were obtained by 48K liquid phase hybridization capture probes technique, and phenotypes were evaluated in multi-environment trails. QTL mapping was performed using the inclusive composite interval mapping (ICIM) method. The RRBLUP model was used for genomic selection to explore the effects of training population size and the number of SNP markers on prediction accuracy. The results showed that the heritabilities of tassel length, tassel primary branch number, tassel secondary branch number, and tassel branch number were 0.82, 0.88, 0.84, and 0.88, respectively. Two QTL, located in bins 1.03 and 4.09, were detected for tassel length with phenotypic variation explained (PVE) of 6.02% and 11.10%, respectively. Two QTL, located in bins 1.05 and 4.05, were detected for tassel primary branch with PVE of 9.17% and 11.75%, respectively. Two QTL, located in bins 2.03 and 3.06, were detected for the tassel secondary branch with PVE of 5.51% and 5.65%, respectively. Two QTLs, located in bins 1.04 and 4.05, were detected for the tassel branch number with PVE of 9.37% and 10.83%, respectively. Tassel primary branch number and tassel branch number identified a same QTL located in bin 4.05 with multiple effects. The prediction accuracy of five-fold cross-validation for genomic selection was 0.36, 0.41, 0.28, and 0.37, respectively. When the training population reached 60% of the total population and the marker density reached 500, a high prediction accuracy could be obtained. [ABSTRACT FROM AUTHOR]