Composition of the final core collection.

Kam Sweet Rice (KSR), a distinctive group of glutinous rice landraces, has evolved over millennia through agro-ecological adaptation by the Dong ethnic group in the 'He' cultivation zone of Southeast Guizhou, China. This study examined the genetic diversity of 388 glutinous rice landraces from the region, comprising 325 KSR and 63 non-KSR varieties, using Simple Sequence Repeat (SSR) sequencing. Results revealed that non-KSR germplasm exhibited significantly higher genetic diversity than KSR germplasm. Collectively, diversity patterns were strongly shaped by the numerical predominance of genetically similar KSR germplasms, resulting in an uneven distribution of genetic diversity between KSR and non-KSR groups. Five strategies were applied to construct and evaluate core collections (see Methods for full details). Among them, the simulated annealing algorithm (SA)-based Allelic Richness Maximization Strategy (SANA) (20% sampling intensity) demonstrated superior performance in preserving genetic diversity, except for the number of alleles (Na) and observed heterozygosity (Ho), where the Modified Heuristic Sampling (M-HS) strategy (13.66% sampling intensity) performed better at lower sampling intensities. By optimizing both approaches, a core collection of 65 germplasms was established, capturing 90.86% of alleles and retaining key genetic parameters. This core set effectively represents the genetic diversity of the entire collection, providing a strong foundation for future germplasm innovation and utilization.