Near-infrared
spectroscopy (NIRS) technology
Near-infrared spectroscopy (NIRS) technology has been proposed as
a rapid analytical method to analyze incoming feed ingredient shipments
in order for the feed formulator to more precisely meet feed nutrient
specifications. Degussa Corporation has been involved in developing
NIRS calibrations for the feed industry for several years. A recent
technical paper in Feedstuffs compared compositional values for
a soybean meal dataset analyzed by classical wet chemistry methods
to their NIRS calibration. One hundred and eighty soybean meal samples
and eighty full-fat soybean meal samples were used in the development
of the data comparing the two analytical methods. The following
compositional data was obtained for soybean meal: |
| |
Classical
Wet-
Chemical Method |
Prediction
Using
NIRS Calibrations |
| Dry Matter |
90.7
|
± |
1.63(*) |
89.8
|
± |
1.62 |
| Crude Protein |
43.8 |
± |
5.47 |
43.9 |
± |
5.43 |
| Methionine |
0.58 |
± |
0.07 |
0.58 |
± |
0.07 |
| Cysteine |
0.66 |
± |
0.08 |
0.66 |
± |
0.07 |
| Methionine & Cysteine |
1.24 |
± |
0.15 |
1.24 |
± |
0.14 |
| Lysine |
2.67 |
± |
0.32 |
2.65 |
± |
0.32 |
| Threonine |
1.70 |
± |
0.21 |
1.70 |
± |
0.20 |
| Tryptophan |
0.59 |
± |
0.07 |
0.59 |
± |
0.08 |
| Arginine |
3.22 |
± |
0.43 |
3.22 |
± |
0.42 |
| Isoleucine |
1.97 |
± |
0.25 |
1.97 |
± |
0.25 |
| Leucine |
3.32 |
± |
0.42 |
3.32 |
± |
0.41 |
| Valine |
2.07 |
± |
0.26 |
2.07 |
± |
0.26 |
| Histidine |
1.17 |
± |
0.14 |
1.18 |
± |
0.26 |
| Phenylalanine |
2.21 |
± |
0.29 |
2.22 |
± |
0.28 |
|
|
| (*) Mean and standard
deviation |
|
The research group also presented charts showing the regression
of the NIRS values to the wet chemistry values. The slope of the
regression could be calculated with the following correlations: |
 |
The significance of this report is that it is possible to use NIRS
equipment to rapidly analyze soybean meal samples for crude protein,
and important amino acids with accuracy similar to traditional wet
chemical methods. The next step is for other laboratories to develop
robust calibrations that can be used on their NIRS equipment.
Note-It should be pointed out that the
standard deviations for the various amino acids are extremely small
in both methods of analysis (wet chemistry and NRIS). The researchers
developed these data from a pool of 260 soybean meal samples collected
from U.S., Brazil and Argentina. The larger standard deviation for
crude protein can be easily justified due to the pooling of data
from hi-pro dehulled soybean meal and meals containing low protein
with hulls. The consistency of soybean meal is a big technical advantage
for soybean meal in formulating livestock and poultry feeds.
Schirmer, B., J. Goodson and J. Fontaine. 2006. Amino acid
predictions by NIRS prove accurate. Feedstuffs, November 7, 2006.
pages 14-15.
|
