Association mapping of mesocotyl and coleoptile length in rice using various genome-wide association study models

The mechanized direct seeding of rice (Oryza sativa L.) is a major trend nowadays. The elongation of rice mesocotyl and coleoptile can facilitate the rapid emergence of seedlings under deep mechanized sowing. Currently, most of the cultivated rice accessions have short mesocotyls or coleoptiles, with only a few related genes cloned. However, understanding and enhancing the ability of rice seedlings to rapidly emerge from deep sowing depths is crucial. Herein, we assessed 745 core rice germplasm accessions sown under a soil cover depth of 10 cm and found few long mesocotyl and coleoptile germplasms. We conducted genome-wide association study using six models to obtain three or more multi-model co-localization candidate regions and calculated Fst between the phenotypes of extreme samples to determine genetic differences. The candidate regions associated with mesocotyl and coleoptile lengths were identified by integrating Fst and multi-model localization results. This multi-model localization method may accelerate the mining of genes related to the mesocotyl and coleoptile, providing valuable targets for functional validation and marker-assisted selection in rice breeding programs. We utilized diverse GWAS models to co-locate candidate loci for mesocotyl and coleoptile elongation. We calculated Fst in phenotypically divergent materials to refine candidate regions. This multi-model localization method provides valuable targets for rice breeding and mechanized cultivation. Rice requires longer mesocotyl and coleoptile for rapid seedling emergence in direct seeding. It is necessary to identify key genes that control mesocotyl and coleoptile traits to facilitate the breeding of new varieties. We analyzed the mesocotyl and coleoptile lengths of 745 rice samples and conducted genome-wide association study using six different models to co-locate candidate loci. Several valuable candidate regions were consistently detected using different models. These regions included previously reported genes related to mesocotyl/coleoptile elongation and other agronomic traits such as plant height, heading date, and yield. Our findings validated the accuracy of the multi-model analysis and provided guidance on efficiently mining genes involved in mesocotyl/coleoptile elongation from those known genes controlling plant growth and development. This study may contribute to the breeding of new rice varieties suitable for direct-seeding by synergistic selection of multi-traits.