Rice is a global staple crop, being the main calorific component of many people living subsistence livelihoods. Rice can accumulate toxic elements such as arsenic and cadmium, and for these elements especially, the crop water management strongly affects uptake. While a number of genes have been identify that regulate the accumulation of arsenic and cadmium in rice grains, these genes have been through the use of mutant studies, over expression or gene knockouts. While the identification of these genes have been instrumental in upstanding the processes of arsenic and cadmium accumulation in rice, there is still scope to utilise natural variation for these traits in breeding programmes. This study was developed to assess genotype by water management interaction, identify quantitative trait loci (QTL) for arsenic and cadmium accumulation and propose candidate genes for lowering grain arsenic and cadmium.

This study utilises the Bengal and Assam Aus Panel to conduct genome wide association (GWA) mapping for grain arsenic and cadmium in shoots and grains of grown over two years under continually flooded (CF) and alternate wetting and drying (AWD). AWD significantly reduced grain arsenic across all cultivars on average by 15.7% and 15.1% in year 1 and 2. AWD significantly increased grain cadmium across all cultivars on average by 50.0% and 92.4% in year 1 and 2 respectively compared to the plants grown under CF. There was strong cultivar by treatment interaction for grain cadmium but it was weaker for arsenic. All traits were strongly influenced by cultivar. GWA mapping identified a large number of QTLs for both arsenic and cadmium traits, with respectively six and- 26 QTLs showing stability across years and/or water treatments. Six of the loci were investigated in detail using an approach of clustering cultivars that had similar haplotypes for the QTL regions and then looking at the phenotypic values across the clusters. A number of the identified QTLs co-localised with known genes involved in arsenic (Lsi2) and cadmium (OsNRAMP5) accumulation. This study has identified a number of novel QTLs for arsenic and cadmium accumulation, as well as cultivars that consistently accumulate less of these elements over multiple field traits. The use of a haplotype clustering approach after genome wide association mapping has allowed for the affect, in terms of elemental accumulation, to be determined for cultivars that share similar genomic sequence. These haplotypes can be further investigated for their potential in breeding low arsenic and cadmium rice cultivars by utilising natural variation.