Effect of different heat treatments during processing on nutrient digestibilities of soybean meal in growing swine

Abstract

Four soybean meals, classified as under, normal, over, and ruminant meals, were processed in a commercial solvent-extraction plant to represent the widest possible range in heat treatment that commercial soybean meals would receive. Under and normal meals were processed through a Crown Iron Works three-compartment desolventizer-toaster (D-T) unit but steam was shut-off in the second compartment when the under meal was processed. The over and ruminant meals were further processed through a four-compartment reject toaster. Retention times in the reject toaster were approximately 33 (over) and 66 min (ruminant). Arginine, Lys, and Cys decreased more than other amino acids with increasing heat treatments (about 7% lower in ruminant than under meal). Trypsin inhibitor (TI, mg/g as-fed sample), urease activity (UA, (DELTA) pH), protein dispersibility index (PDI, %), and nitrogen solubility index (NSI, %) for the under, normal, over, and ruminant meals were 5.3, 4.0, 1.6, 1.8; .19, .11, .01, .02; 54.0, 63.7, 40.6, 7.3, and 27.8, 25.1, 12.5, 7.0, respectively. Nutrient digestibilities of the four meals were determined in a 4 x 4 Latin Square designed trial using barrows (about 25 kg initially) fitted with a simple T-cannula near the terminal ileum. An additional fifth period was included to determine the digestibility of flash-desolventized raw soyflakes that were not processed through either the D-T or reject toaster (TI, UA, PDI, and NSI wre 27.5, 2.04, 80.6, 75.7, respectively). Apparent digestibilities of dry matter, N, gross energy, and individual amino acids were similar for the four heat-treated meals both at the end of the small intestine and over the total tract. Nitrogen, Lys, Trp, and Thr ileal digestibilities for the under, normal, over, and ruminant meals were 82.2, 81.6, 80.8, 80.7; 87.6, 85.2, 84.9, 82.6; 78.0, 79.1, 78.4, 79.2, and 77.9, 75.0, 76.5, 76.6, respectively. Digestibilities for the raw soyflakes were much lower and differences were greater at the terminal ileum than over the total tract. Four in vitro enzyme digestion assays (pepsin, pepsin-trypsin, pepsin-pancreatin, and pepsin-pancreatic juice) were performed on the five soybean products of the present study plus 14 other soybean products, four cottonseed meals, and one meat and bone meal with known swine in vivo digestibilities. The most effective assay to predict ileal N digestibility when all products or when only soybean products were included was the pepsin-pancreatic juice method (R('2) = .58, .65, respectively, P < .01). None of the in vitro assays was highly correlated with ileal N digestibilities of the high-protein feedstuffs.