Application of Field Scale High-Throughput Phenotyping for Gene Discovery and Breeding in Rice

Improving crop yields through breeding requires genetic variation and precise prediction of a genotype's fit to a target environment.  Next generation breeding approaches place an ever increasing demand for large field scale approaches for phenotyping.  At IRRI, we have developed a tractor-based High-Throughput Phenotyping (HTP) system that uses proximal sensors for monitoring multispectral reflectance, canopy temperature, and plant height that accurately predict manually-measured phenotypes.  Four populations have now been evaluated using HTP including: a large set of recombinant inbred lines (1,700 IR64 x Aswina RILs), a diversity panel of 300 Indica accessions (PRAY), a multiparental intercross population (MAGIC), and a drought screening panel (300 IR64 x Dular RILs). Results show high genotypic correlation between HTP traits and manually collected traits measured throughout the season and at harvest. HTP traits - particularly different spectral indices - were correlated with time to flowering, plant height, biomass, tiller number, nitrogen content, and yield.    Initial results show co-localization of QTL detected by HTP and manually measured biomass.  The efficient and non-destructive nature of HTP provides the unique ability to monitor dynamic plant responses such as plant growth, development, and response to stress. These results suggest that HTP will be useful for association mapping of QTL for important agronomic traits. Drone-based technologies are now being tested in conjunction with the tractor-based system for the development of more efficient and systematic screening of large breeding populations and multi-environment testing of advanced breeding materials.