Table 4.

Validation partial least square regression (PLSR) models for trigonelline and 4-hydroxy isoleucine.

Traits	Trigonelline				4-hydroxyisoleucine
	R2	RMSEP	P-value	VIP	R2	RMSEP	P-value	VIP
Aspartic acid	−0.12	0.052	0.886	0.018	−0.15	22.92	0.644	0.78
Glutamic acid	−0.07	0.050	0.486	0.03	−0.11	22.8	0.785	0.61
Valine	−0.085	0.050	0.903	0.007	−0.14	22.16	0.777	0.70
Phenylalanine	−0.052	0.050	0.304	0.07	−0.067	21.87	0.655	0.24
Isoleucine	0.314	0.048	0.059	0.03	0.32	22.16	0.0008***	1.74
Leucine	−0.14	0.047	0.979	0.041	−0.197	22.30	0.939	0.031
Antioxidant	0.087	0.046	0.477	0.87	0.062	23.96	0.0385*	1.20
Flavonoid	−0.064	0.044	0.188	1.04	−0.066	24.68	0.86	0.31
Phenol	−0.087	0.043	0.288	0.47	−0.104	23.9	0.730	0.32
Lysine	−0.082	0.043	0.439	0.28	−0.112	23.40	0.390	0.33
Pod number	−0.048	0.041	0.723	0.07	−0.083	22.77	0.216	0.41
Seed per pod	−0.009	0.041	0.063.	1.31	0.025	21.72	0.261	0.87
Pod length	−0.0511	0.041	0.131	1.14	−0.040	20.74	0.536	0.56
Seed weight	0.130	0.040	0.037*	1.88	0.220	21.00	0.048*	1.32
Pod width	−0.063	0.040	0.195	0.83	−0.066	21.28	0.894.	0.31
Seed length	−0.037	0.040	0.599	0.040	0.080	20.11	0.145	0.75
Seed width	0.061	0.040	0.033*	1.38	0.11	20.49	0.078	1.13
Solar	0.30	0.040	0.194	1.32	0.30	22.38	0.0002***	1.76
Temperature	0.26	0.040	0.593	1.02	0.23	22.91	0.0006***	1.77
Rain	0.146	0.040	0.841	0.93	0.079	22.98	0.0045**	1.74

Variables importance in the projection (VIP) greater than one (bold) were used to select relevant latent variables. Root mean square errors (RMSEP) and R² were applied for cross-validation in PLSR analysis.

Significance. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.