Abstract
<jats:p>Adapting crops to climate change and resource constraints requires leveraging genotype-by-environment interactions (G×E) and the polygenic control of early stress responses. Dry direct seeding (DSR) is an environmentally sustainable rice cultivation method that reduces water use, labor inputs, and methane emissions. However, it increases susceptibility to iron deficiency-induced chlorosis during establishment. Here, we dissect the genetic architecture of Fe-deficiency tolerance across three developmental stages and two environments using the diverse Oryza sativa 3K mini-core panel (n ≈ 372) evaluated under DSR conditions. We phenotyped plants under control and Fe-deficient conditions at 15, 28, and 45 days after sowing (DAS) using three phenotyping approaches: field visual rating, SPAD-based chlorophyll content, and a computer vision (CV) phenotyping pipeline that generated quantitative chlorosis severity scores and captured within-canopy heterogeneity. Our CV phenotypes are both correlated with and complementary to the more traditional phenotypes. We conducted a temporal genome-wide association study and haplotype analyses that revealed stage- and environment-dependent genetic effects. The identified loci were primarily time- and environment-specific, with significant effects in one environment but none in the other (conditional neutrality). At 28 DAS, when control and Fe-deficient plants were most differentiated, we identified a locus with stable effects on the iron-deficiency response across environments. This stable genetic signal is linked to the RCK locus (LOC_Os02g40450), which harbors a superior gene haplotype that consistently reduces chlorosis severity across environments. We identified seventeen additional candidate genes (OsAVP1, OsPRR1, WRKY62, IIP4, OGR1, OsPCBP, HWH1, OsCKI1, and MYB45, (MRLK47, MRLK48, UAA, LRR-RLK, OsRR8, WRKY62, Fbox-14, SEU, OsPCBP, and OGR1) that showed time- and environment-specific haplotype superiority, with most tolerant haplotypes enriched in Aus subpopulation backgrounds. These results highlight the shifting genetic architecture of Fe-deficiency tolerance and identify potentially functional candidate gene haplotypes and stable genetic signals critical for breeding rice adapted to DSR systems.</jats:p>