Large scale wheat data integration improves genomic prediction accuracy with the potential to facilitate international breeding collaborations

Effective data integration across diverse sources is essential for maximising genetic gain and accurately predicting plant performance in modern agriculture. However, individual breeding programs often face significant logistical and financial barriers when attempting to test large, diverse populations across a wide range of global environments. Here we show that synchronising genomic data from two prominent wheat breeding programs with a total of 11,609 wheat accessions that were evaluated across 79 environments establishes a robust platform that significantly improves genetic prediction accuracy and the power to identify complex trait associations. By developing a computationally efficient statistical model, we demonstrate that combining these massive datasets increases prediction accuracy by up to 13% while drastically reducing the time and computational resources required for analysis. This collaborative approach leverages existing investments and taps into broader genetic diversity to overcome the geographical and resource limitations of isolated programs. Our findings highlight the transformative power of collective data integration and provide a practical framework to revolutionise current breeding strategies. Ultimately, our framework has the potential to facilitate international cooperation and streamline the development of climate-resilient varieties, thereby facilitating more sustainable agricultural practices and strengthening global food security.