Research was conducted to develop a mathematic model that estimates the value of soybean meal based on soybean composition.  The research involved collecting 8,282 soybean samples from 29 states during 2013-2016 and completing nutrient analyses of the individual samples.  A least-cost typical mid-finishing swine diet was formulated and used to develop relative comparisons of the individual soybean samples.  The researchers assumed constant energy contents and the meal values would be primarily based on the protein and amino acids differences in the individual soybean samples.  The model developed was:   

Soybean Meal Value ($) per ton equals 72.16361 + 1.65587 x Protein + 27.24420 x Lysine + 57.25507 x Methionine + 84.56061 x Tryptophan + 21.80598 x Isoleucine.  (The units of the independent variables are expressed as percent on a dry weight basis).

The predictive reliability of the model was evaluated with independent soybean samples and found to have a correlation of R²=0.99 between calculated and actual soybean meal values.  There is a correlation R²=0.82 between soybean seed protein and soybean meal value.  The predictive ability of the formula was further improved by including lysine, methionine, tryptophan and isoleucine values.

The study showed no correlation (R²=0.01) between soybean yield and soybean protein content or calculated soybean meal value.  They found ample variation between calculated soybean meal values between soybean production regions and within states.