Nature Genet. http://doi.org/bmjp (2016)

Next-generation sequencing greatly facilitates the application of genome-wide association studies (GWAS) in identifying agronomically important genes. However, population stratification and long linkage disequilibrium (LD) still impede the effective implementation of GWAS. To improve the power of GWAS, Kenji Yano, at Nagoya University, Japan and colleagues have developed a new method for agronomic gene identification.

The researchers carefully selected a panel of 176 japonica rice varieties with low structure and the typically long LD of rice populations. GWAS analysis using a linear mixed model detected 26 loci associated with heading date.

For five of the loci, the candidate region was estimated using pair-wise LD correlations. ‘Large-effect’ polymorphisms that are non-synonymous or disrupt splicing junctions were then specifically analysed to rapidly narrow down candidate alleles. This strategy allowed the authors to identify one gene in four of the five loci, whose role in heading date control was previously known or validated by transgenic experiments by the authors. Another gene, NAL1, which controls panicle number, was also discovered in a similar manner.

This approach failed to demonstrate statistical significance for the Hd1 gene underlying the fifth heading-date locus, which the researchers attributed to allelic heterogeneity. Nevertheless, gene-based association analysis could be a solution for allelic heterogeneity through mapping multi-allelic genes.