Ridge regression and deep learning models for genome-wide selection of complex traits in New Mexican Chile peppers

Table 1 Mean prediction accuracies across the different Bayesian ridge regression and BLUP and deep learning models for yield and agronomic traits in chile pepper using whole marker and a subset (s) of SNP loci for genomic selection

¹FPD- First pod date (0.73); FT- Flowering time (0.73); GRN- Mature green yield (0.58); PHT- Plant height (0.61); PWDTH- Plant width (0.41); RED- Mature red yield (0.20); TPW- Ten pod weight (0.88); TYP- Total yield per plant (0.20). Values in parentheses are broad-sense heritability (H²) for the traits as reported by Lozada et al. [30].
² Mean prediction accuracy significantly different with the accuracy for whole genome marker data at P < 0.05 (P = 0.045; Student t-test).
³ Mean prediction accuracy significantly different with the accuracy for whole genome marker data at P < 0. 05 (P = 0.0115; Student t-test).
⁴ Mean prediction accuracy significantly different with the accuracy for whole genome marker data at P < 0. 001 (P = 0.0006; Student t-test).
⁵ Mean prediction accuracy significantly different with the accuracy for whole genome marker data at P < 0. 05 (P = 0.009; Student t-test).

ISSN: 2730-6844