Full-fat soybean origin and thermal processing affect protein secondary structure, in vitro digestion kinetics, and predicted standardized ileal amino acid digestibility in broiler chickens

BACKGROUND: This study evaluated the chemical composition, protein secondary structure, crude protein (CP) solubilization, and in-vitro amino acid (AA) digestibility of three Kenyan full-fat soybean (FFSB) samples (S1, S2, S3) with unknown processing histories and compared with Canadian samples: raw FFSB, extruded FFSB at 160.0 °C, and FFSB roasted at 136.7 °C, 151.7 °C, or 154.4 °C. Protein structure was characterized using attenuated total reflectance-Fourier transform infrared spectroscopy. A two-phase in-vitro digestion model was used to determine CP solubilization and AA digestibility. Observed in-vivo standardized ileal digestibility (SID) values from a study using FFSB roasted at 136.7 °C were regressed on corresponding in-vitro AA digestibility values to generate predicted SID values. RESULTS: Thermal processing reduced trypsin inhibitor activity (TIA) in Canadian FFSB from 17.42 TIU/mg in raw FFSB to 3.85 TIU/mg in extruded FFSB and to 4.77, 4.59, and 2.30 TIU/mg in samples roasted at 136.7, 151.7, and 154.4 °C, respectively. In contrast, Kenyan Samples 1, 2, and 3 had residual TIA values of 8.00, 6.00, and 12.00 TIU/mg, respectively. Protein secondary structure differed among samples (P < 0.001), with the 136.7 °C-roasted sample having greater aggregated β-sheet and β-turn contents and lower random coil content than most samples. CP solubilization differed among samples at each digestion time point (P < 0.001). At the end of the intestinal phase, processed Canadian samples had greater CP solubilization than raw FFSB but did not differ from one another, whereas Kenyan Samples 2 and 3 were comparable to Canadian raw FFSB. In-vitro AA digestibility differed among samples for all AA (P ≤ 0.001). For mean total AA digestibility, the 154.4 °C and 151.7 °C-roasted sample had the highest digestibility; Kenyan Sample 1 was greater than Samples 2 and 3, and Sample 3 did not differ from raw FFSB. The regression model showed agreement at the mean level but larger prediction errors for lysine, aspartic acid, and proline, indicating AA-specific limitations. CONCLUSIONS: Overall, processing conditions and ingredient origin influenced FFSB quality, with roasting at 151.7-154.4 °C producing favorable structural and digestibility responses. The variability observed among Kenyan samples highlights the need for routine quality assessment of commercial FFSB before feed formulation.