Transcriptome-Based Prediction of Heterosis and Hybrid Performance

Abstract

Heterosis is defined as the ability of hybrids to outperform their parents with respect to various characteristics and agronomical important traits (Shull 1948). Various plant traits and particularly pronounced yield display heterosis. Hybrid breeding is based upon the phenomenon and prediction of hybrid performance and heterosis are important applications to increase the efficiency of hybrid breeding programs. Traditional phenotypic evaluation is still the common methodology to estimate general combining abilities for the predictions of hybrid traits (Choudhary et al. 2008). Molecular markers like RFLPs (restriction fragment length polymorphism), RAPDs (randomly amplified polymorphic DNA) or AFLPs (amplified fragment length polymorphism) can be employed to improve the prediction of complex traits like yield. An example is the application of AFLP-Markers in a linear regression approach to predict hybrid performance and SCA in intergroup-crosses of maize (Vuylsteke et al. 2000). For a long time molecular marker-based genetic distances has not proven to be a reliable approach for the prediction of hybrid characteristics in crops (Melchinger 1999) and might have limitations for prediction, as e.g. in Arabidopsis the correlation between heterosis and genetic distance was not significant (Meyer et al. 2004; Stokes et al. 2007). However, the improvement of DNA marker based approaches and its application to factorial crosses in maize led to a prediction accuracy that is comparable to phenotype based estimates (Schrag et al. 2006, 2007). To further increase prediction abilities for hybrid breeding, various molecular compounds of the plant, like DNA methylation states, transcripts, proteins and metabolites, are currently considered and tested as new markers with diagnostic and predictive potential. Current results related to specific compounds are discussed in the various chapters of this section. Here we address the development and applicability of RNA expression data for the prediction of heterosis and hybrid performance.