Transcriptome-based distance measures for grouping of germplasm and prediction of hybrid performance in maize

A unifying theory for general multigenic heterosis: energy efficiency, protein metabolism, and implications for molecular breeding

Hybrids between genetically diverse varieties display enhanced growth, and increased total biomass, stress resistance and grain yield. Gene expression and metabolic studies in maize, rice and other species suggest that protein metabolism plays a role in the growth differences between hybrids and inbreds. Single trait heterosis can be explained by the existing theories of dominance, overdominance and epistasis. General multigenic heterosis is observed in a wide variety of different species and is likely to share a common underlying biological mechanism. This review presents a model to explain differences in growth and yield caused by general multigenic heterosis. The model describes multigenic heterosis in terms of energy-use efficiency and faster cell cycle progression where hybrids have more efficient growth than inbreds because of differences in protein metabolism. The proposed model is consistent with the observed variation of gene expression in different pairs of inbred lines and hybrid offspring as well as growth differences in polyploids and aneuploids. It also suggests an approach to enhance yield gains in both hybrid and inbred crops via the creation of an appropriate computational analysis pipeline coupled to an efficient molecular breeding program.

Enriched partial correlations in genome-wide gene expression profiles of hybrids (A. thaliana): a systems biological approach towards the molecular basis of heterosis

Heterosis is a well-known phenomenon but the underlying molecular mechanisms are not yet established. To contribute to the understanding of heterosis at the molecular level, we analyzed genome-wide gene expression profile data of Arabidopsis thaliana in a systems biological approach. We used partial correlations to estimate the global interaction structure of regulatory networks. Our hypothesis states that heterosis comes with an increased number of partial correlations which we interpret as increased numbers of regulatory interactions leading to enlarged adaptability of the hybrids. This hypothesis is true for mid-parent heterosis for our dataset of gene expression in two homozygous parental lines and their reciprocal crosses. For the case of best-parent heterosis just one hybrid is significant regarding our hypothesis based on a resampling analysis. Summarizing, both metabolome and gene expression level of our illustrative dataset support our proposal of a systems biological approach towards a molecular basis of heterosis.

Prediction of maize single-cross hybrid performance: support vector machine regression versus best linear prediction

Accurate prediction of the phenotypic performance of a hybrid plant based on the molecular fingerprints of its parents should lead to a more cost-effective breeding programme as it allows to reduce the number of expensive field evaluations. The construction of a reliable prediction model requires a representative sample of hybrids for which both molecular and phenotypic information are accessible. This phenotypic information is usually readily available as typical breeding programmes test numerous new hybrids in multi-location field trials on a yearly basis. Earlier studies indicated that a linear mixed model analysis of this typically unbalanced phenotypic data allows to construct epsilon-insensitive support vector machine regression and best linear prediction models for predicting the performance of single-cross maize hybrids. We compare these prediction methods using different subsets of the phenotypic and marker data of a commercial maize breeding programme and evaluate the resulting prediction accuracies by means of a specifically designed field experiment. This balanced field trial allows to assess the reliability of the cross-validation prediction accuracies reported here and in earlier studies. The limits of the predictive capabilities of both prediction methods are further examined by reducing the number of training hybrids and the size of the molecular fingerprints. The results indicate a considerable discrepancy between prediction accuracies obtained by cross-validation procedures and those obtained by correlating the predictions with the results of a validation field trial. The prediction accuracy of best linear prediction was less sensitive to a reduction of the number of training examples compared with that of support vector machine regression. The latter was, however, better at predicting hybrid performance when the size of the molecular fingerprints was reduced, especially if the initial set of markers had a low information content.

Prediction of hybrid biomass in Arabidopsis thaliana by selected parental SNP and metabolic markers

A recombinant inbred line (RIL) population, derived from two Arabidopsis thaliana accessions, and the corresponding testcrosses with these two original accessions were used for the development and validation of machine learning models to predict the biomass of hybrids. Genetic and metabolic information of the RILs served as predictors. Feature selection reduced the number of variables (genetic and metabolic markers) in the models by more than 80% without impairing the predictive power. Thus, potential biomarkers have been revealed. Metabolites were shown to bear information on inherited macroscopic phenotypes. This proof of concept could be interesting for breeders. The example population exhibits substantial mid-parent biomass heterosis. The results of feature selection could therefore be used to shed light on the origin of heterosis. In this respect, mainly dominance effects were detected.

Correlation between parental transcriptome and field data for the characterization of heterosis in Zea mays L

Heterosis is widely exploited in plant breeding, although its molecular basis is still not fully understood. For the characterization of this phenomenon and the development of transcriptome-based methods to predict hybrid performance (HP), we applied a microarray (46k) analysis of 21 European maize (Zea mays L.), 14 dent and 7 flint parental inbred lines. Expression profiles of the parental inbreds at the seedling stage were correlated with grain yield (GY) and grain dry matter content (GDMC) of 98 flint x dent factorial crosses at six locations. We observed highly significant correlations of the parental expression levels of certain differentially expressed genes with heterosis and HP for GY and also with HP for GDMC. This strong correlation provided first evidence toward a prediction potential of the genes and their expression levels. The identified gene set based on the parental transcriptome data revealed functional characteristics of HP and heterosis. Gene ontology (GO) analyses were performed to compare genes correlated with their expression pattern to HP for GY and GDMC, respectively. Between these gene groups, mostly different functional classes of genes were found to be enriched or underrepresented. The phenomenon of heterosis was characterized by the over- and underrepresentation of specific GO terms among heterosis-correlated genes.

Background correction of two-colour cDNA microarray data using spatial smoothing methods

The analysis of two-colour cDNA microarray data usually involves subtracting background values from foreground values prior to normalization and further analysis. This approach has the advantage of reducing bias and the disadvantage of blowing up the variance of lower abundant spots. Whenever background subtraction is considered, it implicitly assumes locally constant background values. In practice, this assumption is often not met, which casts doubts on the usefulness of simple background subtraction. In order to improve background correction, we propose local background smoothing within the pre-processing pipeline of cDNA microarray data prior to background correction. For this purpose, we employ a geostatistical framework with ordinary kriging using both isotropic and anisotropic models of spatial correlation and 2-D locally weighted regression. We show that application of local background smoothing prior to background correction is beneficial in comparison to using raw background estimates. This is done using data of a self-versus-self experiment in Arabidopsis where subsets of differentially expressed genes were simulated. Using locally smoothed background values in conjunction with existing background correction methods increases the power, increases the accuracy and decreases the number of false positive results.