We studied the effects of harvesting year and country of origin of the beans on the chemical composition, amino acid (AA) profile, and protein quality indicators of commercial soybean meals (SBM). In total, 312 samples of SBM from Argentina (ARG; n = 98), Brazil (BRA; n = 109), and United States of America (USA; n = 105) were collected at random from 2016 to 2022 (n = 55, 62, 27, 51, 34, 46, and 37, in each consecutive year) in crushing plants and feed mills at different locations of the European Union by specialized personnel. All chemical analyses were conducted by wet chemistry, except AA that were determined by NIRS. The data were analyzed as a completely randomized design with 21 treatments arranged as a 7 x 3 factorial with harvesting year and country of origin of the beans as main effects, using the MIXED procedure of SAS. The interactions between year crop and country of origin of the beans were significant for all the variables studied (P < 0.05) except EE. On 88% DM bases, harvest year affected all the variables on chemical composition and protein quality indicators of the SBM (P < 0.05). Independently of the country of origin of the beans, the average values per year for each variable ranged from 5.72 to 7.36% for sucrose, 45.9 to 47.3% for CP, 6.06 to 6.13% for Lys profile (% CP), and 1.79 to 2.64 mg/g for TIA (P < 0.001). Independently of the year of harvest, the moisture content of the ARG meals was lower than that of the BRA and the USA meals (11.2 vs. 11.6 vs. 11.8%; P < 0.01). BRA meals had more protein (47.3 vs. 46.1 vs. 46.1%) and raffinose (1.36 vs. 1.36 vs. 1.19%) but less sucrose (5.73 vs. 7.16 vs. 7.01%) and stachyose (4.22 vs. 4.64 vs. 4.91%) than the ARG and USA meals (P < 0.001). However, no differences among soybean origins were detected for EE or NDF. The content and profile of the AA, varied with the country of origin of the beans. In this respect, Lys profile (% CP) was better for the USA meals than for ARG and BRA meals (6.13 vs. 6.10 vs. 6.07%; P < 0.001). All protein quality indicators, except UA, were higher for the USA than for the South American meals (P < 0.001).
It is concluded that the chemical composition, AA profile, and protein quality indicators of the SBM vary with the harvest year and the country of origin of the meals. Consequently, feed mill managers should use different matrices to estimate the chemical composition and nutritive value of SBM of different origins, taking into account the inherent variability of its nutrient content due to these two variables. AMINONIR NIRS estimates of FFS and SBM TIA can be predicted from NIRS estimates of RS with moderate accuracy, but predictions in SBM were not accurate.