Context. Establishment of true heterotic pattern in maize germplasm can increase the efficiency of hybrid breeding. Heterosis is dependent on the genetic diversity of parents and the extent of dominance at different loci. Estimation of genetic diversity through use of molecular markers is routine practice in maize breeding.

Aims. The present study was designed to test whether simple sequence repeat (SSR) markers linked to yield-contributing traits are more reliable for heterotic grouping than random SSRs.

Methods. Diallel crosses developed among 19 inbred lines were evaluated at multi-locations. The genotypes were also grouped using polymorphic random (50) and linked (47) SSRs.

Key results. The crosses generated with lines belonging to different heterotic groups of linked SSR markers did not reveal any superiority over the crosses of the diallel set. By contrast, mean performance of inter-heterotic group crosses generated on the basis of random markers was superior to that of intra-heterotic crosses. Specific combining ability effects did not reveal any significant association with genetic distance of random or linked markers.

Conclusions. The lack of improved efficiency of linked markers over random markers can be attributed to factors including the quantitative nature of the trait, genotype × environment interactions, genetic background of germplasm in which the markers are expressed, and multiple alleles.

Implications. Markers linked to yield-contributing traits are no more reliable for heterotic grouping than random markers.