Genetic Improvement of Rice for Yield, Grain Quality and Abiotic Stress Tolerance: An Integrative Critical Review

Abstract

Rice improvement now operates at the intersection of three objectives that were historically treated more independently: raising yield, maintaining or improving grain quality, and stabilising performance under abiotic stress. This critical narrative review evaluates the genetic architecture, breeding evidence and translational maturity of these objectives in Asian rice (Oryza sativa L.), while drawing selectively on African rice and wild Oryza germplasm where they illuminate useful alleles or mechanisms. Literature published from 2000 to 5 August 2026 was prioritised, with earlier or foundational work retained when necessary. The synthesis distinguishes large-effect loci that have translated effectively into breeding, such as SUB1A, from traits whose expression is strongly background- or environment-dependent. Yield improvement has benefited from genes controlling stature, panicle architecture, grain number, grain size and heading date, but the same loci can alter maturity, lodging risk, source–sink relations and market-preferred grain dimensions. Grain quality is similarly modular rather than unitary: appearance, milling, cooking, eating, aroma and nutritional attributes arise from partially distinct pathways, with Wx, ALK, BADH2, Chalk5 and grain-size regulators illustrating both tractability and pleiotropic constraint. For abiotic stress tolerance, major loci affecting sodium transport, submergence quiescence, root architecture, anaerobic germination, phosphorus acquisition, heat tolerance and chilling adaptation provide strong mechanistic anchors, yet combined-stress and grain-filling environments remain underrepresented. Genomics, genomic selection, pangenomics and genome editing broaden the accessible allele space, but prediction and editing do not remove the need for rigorous multi-environment phenotyping. The central conclusion is that future genetic gain will depend less on stacking nominally favourable alleles than on designing compatible haplotype combinations whose effects remain stable across target production environments and consumer quality requirements.