PotatoMASH is a cost-effective marker system for genomic prediction in potato based on short-read haplotypes

Amplicon sequencing and read-backed haplotyping enable cost-effective genomic prediction in potato, with SNPs and haplotags showing similar performance across 23 traits, supporting scalable breeding applications and competitive prediction ability with GBS. Genomic prediction (GP) supports plant breeding by accelerating genetic improvement; however, the high cost associated with dense genotyping platforms restricts their use in routine breeding. This study evaluates the efficacy of PotatoMASH, a cost-effective, low-density, amplicon-sequencing platform generating SNPs and short-read multi-allelic haplotypes (haplotags), for GP in potato. First, we compared the prediction ability (PA) achieved using 2,236 SNPs and 2,000-3,390 haplotags from 339 amplicon loci of PotatoMASH with previously reported PA values obtained from a high-density 43.6 k SNP GBS dataset. PA was only moderately reduced, by 14% for SNPs and 9% for haplotags, indicating that the platform offers a scalable alternative to GBS. We then applied it to a diploid panel for GP of 23 agronomic, quality, and morphological traits. Both marker types yielded medium to high PA (0.29-0.81) across the traits. Haplotags outperformed SNPs in 11 traits, while SNPs performed better in six, with no difference in the remaining six traits. We conclude that PotatoMASH, which facilitates the concurrent detection of both SNPs and haplotypes at a reduced cost, provides a versatile and economical genotyping solution suitable for integrated pipelines that combine marker-assisted selection (MAS) and GP in potato breeding.