Heterotic group classification and heterosis prediction in recombinant maize inbred lines

BACKGROUND: Effective hybrid breeding in maize depends on accurate heterotic group classification and reliable prediction of heterosis. However, the reliability of different prediction methods and the factors influencing their performance remain unclear. RESULTS: This study assessed heterotic group classification methods and heterosis prediction models using 102 crosses derived from three testers and 34 recombinant inbred lines (RILs). Among the methods classified, the heterotic group-specific and general combining ability (HSGCA) approach was the most effective for heterotic group classification. The Tropical group contained a higher number of molecular markers in key genomic regions, whereas crosses involving the Tropical × Reid pattern showed a greater proportion of heterozygous markers. Both the Tropical × Reid and Tropical × Non-Reid heterotic patterns exhibited higher genetic distance (GD). Genetic distance showed the strongest correlation with grain yield heterosis among all molecular measures evaluated. After the heterotic group classification, inter-group GD correlated more strongly with heterosis than intra-group GD. The GD within the Tropical × Reid inter-group pattern showed the highest positive correlation with heterosis. CONCLUSIONS: For grain yield heterosis in maize, it is concluded that: (1) HSGCA is the most suitable method for heterotic group classification; (2) GD is a robust molecular predictor of heterosis; and (3) the predictive power of GD is maximized within the Tropical × Reid inter-group heterotic pattern, which offers a clear framework for future breeding programs. © 2026 Society of Chemical Industry.