Low coverage whole genome sequencing is an effective strategy for genome-wide association studies (GWAS) in humans, due to the availability of large reference panels for genotype imputation. However, the utility of this strategy in other species without reference panels is unclear. Using simulations, we show that this approach is even more relevant in inbred species such as rice (Oryza sativa L.), which are effectively haploid, allowing easy haplotype construction and imputation-based genotype calling, even without the availability of large reference panels. We sequence 203 rice varieties with well-characterized phenotypes from the USDA Rice Minicore Collection at an average depth of 1.5× and used the data for mapping three traits. For the first two traits, amylose content and seed length, we show that our approach leads to direct identification of the previously identified causal SNP in the major-effect loci. For the third trait, pericarp color, an important trait selected for during domestication, we discover a new major-effect locus. While known loci can explain color variation in the varieties of two main subspecies of Asian domesticated rice, japonica andindica, the new locus we found is unique to another domesticated rice subgroup, aus and together with existing loci fully explain the major variation in pericarp color in aus. Our discovery of a unique genetic basis of white pericarp in aus, provides an example of convergent evolution during rice domestication and suggests that aus may have a domestication history independent of japonica and indica.