Abstract
Pungency levels (capsaicinoid content) are critical traits influencing the quality and commercial value of chili peppers (Capsicum annuum). However, their complex inheritance patterns make controlling them challenging when crossing different progeny in current breeding programs. As a potential solution, we explored genomic prediction (GP) for crossing different progeny based solely on parental data. In this initial study, we assessed the feasibility of GP in 156 F(1) accessions derived from 20 parents within 132 inbred C. annuum accessions. Capsaicinoid content (capsaicin, dihydrocapsaicin, and their total) was quantified using high-performance liquid chromatography. Inheritance analysis revealed that nearly half of the F(1) accessions exhibited high-parent heterosis (F(1) > higher parent), particularly in crosses between lower-pungency parents. We then performed GP for F(1) accessions using 3,149 single nucleotide polymorphisms from inbred accessions. Among 11 models tested, GBLUP-GAUSS tended to show high accuracy, with predicted values showing a significant positive correlation (r = 0.770, P < 0.01) with observed capsaicinoid content (μg·gDW(-1)), although the involvement of heterosis in reducing accuracy was observed. These findings suggest that GP can effectively rank pungency levels among F(1) progeny based solely on parental information, providing valuable insights for developing GP-based breeding strategies in chili pepper.