A Multi-Gene Matrix for Precision Breeding of Rice Grain Quality

Rice grain quality is governed by a network of interacting loci, which makes the simultaneous improvement of several grading indices difficult. We phenotyped 240 indica accessions for 12 quality traits and graded six core indices—head rice rate, chalkiness degree, transparency, amylose content, gel consistency, and alkali spreading value—against the Chinese national standard NY/T 593-2021. Pairwise correlations resolved the traits into grain-shape, chalkiness, and starch clusters and quantified the trade-offs between them. KASP markers for GS3, GW5, Chalk5, Wx, and ALK were then used to test each locus against the six indices. Head rice rate responded to both chalkiness and grain shape; chalkiness degree was governed mainly by Chalk5 and modified by the grain-shape loci; transparency depended on Wx/ALK and on chalkiness; amylose content and alkali spreading value tracked Wx and ALK, respectively; and gel consistency reflected the Wx–ALK combination. The haplotype GW5/GS3/Chalk5L/Wxb/ALKb reached grade 2 or better on all six indices but occurred in only two of the 240 accessions. Near-isogenic lines carrying GW5/GS3/Chalk5L in the Xibaizhan background produced longer, narrower grains and reduced the chalkiness degree from 17.1% to 5.3%, confirming the contribution of the appearance module within the matrix.