J. Appl. Poult. Res.
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J APPL POULT RES 2008. 17:243-248. doi:10.3382/japr.2007-00078
© 2008 Poultry Science Association
This Article
Right arrow Summary Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lordelo, M. M.
Right arrow Articles by Davis, A. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Lordelo, M. M.
Right arrow Articles by Davis, A. J.

Research Reports

Near Infrared Reflectance Spectroscopy for the Determination of Free Gossypol in Cottonseed Meal1

M. M. Lordelo*, S. A. Shaaban{dagger}, N. M. Dale*, M. C. Calhoun{ddagger}, P. F. Vendrel{dagger} and A. J. Davis*,2

* Department of Poultry Science, The University of Georgia, Athens 30602-2772; {dagger} Agricultural and Environmental Services Laboratories, The University of Georgia, Athens 30602-4356; and {ddagger} Agricultural Research and Extension Center, Texas A&M University, San Angelo, TX 76901-9741

2 Corresponding author: ajdavis{at}uga.edu


   SUMMARY
TOP
 SUMMARY
DESCRIPTION OF PROBLEM
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS AND APPLICATIONS
 REFERENCES AND NOTES
 
Gossypol is a toxic polyphenolic compound produced by the pigment glands of the cotton plant. The free gossypol content of cottonseed meal (CSM) is commonly determined by the American Oil Chemists’ Society (AOCS) wet chemistry method. The AOCS method, however, is laboratory-intensive, time-consuming, and therefore, not practical for quick field analyses. To determine if the free gossypol content of CSM could be predicted by near infrared reflectance spectroscopy (NIRS), CSM samples were collected from all over the world. All CSM samples were ground and a portion of each analyzed for free gossypol by the AOCS procedure (reference data) and by NIRS (reflectance data). Both reflectance and reference data were combined in a calibration. The coefficient of determination (r2) and standard error of prediction (SEP) were used to assess the calibration accuracy. The r2 was 0.728, and the SEP was 0.034 for the initial calibration that included samples from all over the world. However, the r2 and SEP improved to 0.921 and 0.014, respectively, if the calibration was made using CSM samples only from the United States. These results indicate that a general prediction equation can be developed to predict the free gossypol content of CSM by NIRS. From a practical standpoint, NIRS technology provides a method for quickly assessing whether a particular batch of CSM has a free gossypol content low enough to be suitable for use in poultry diets.

Key Words: cottonseed meal • gossypol • near infrared reflectance spectroscopy


   DESCRIPTION OF PROBLEM
TOP
 SUMMARY
 DESCRIPTION OF PROBLEM
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS AND APPLICATIONS
 REFERENCES AND NOTES
 
Cottonseed meal can be an attractive alternative protein source for poultry diets. However, the possibility that it contains high levels of gossypol discourages its use. Gossypol is a toxic polyphenolic compound produced by, and located in, the pigment glands of the cotton plant. The gossypol found in CSM exists in both a free and bound form. Because of its overall negative charge, gossypol tends to form complexes with positively charged molecules such as lysine in CSM [1]. In chickens, free gossypol is assumed to be more readily absorbed from the gastrointestinal tract than bound gossypol and therefore to be the primary cause of observed negative effects associated with its feeding [1–3].

In broiler production, feeding diets that contain CSM with a high level of gossypol has been attributed to poor weight gain [1, 3–6] and poor feed efficiency [6–9]. However, CSM containing low levels of gossypol can successfully be used in broiler diets that contain adequate levels of all essential nutrients, with no adverse effects on body weight or mortality [10, 11]. Furthermore, the lower nutrient density of CSM, when compared with soybean meal, makes it a desirable protein source for broiler breeder pullets [12]. In fact, utilizing CSM in broiler breeder pullet diets reduces the need for severe feed restriction, which results in better flock body weight uniformity [12].

Although CSM can be successfully utilized in poultry diets, it typically is not included because the gossypol content is unknown. The gossypol concentration of CSM can vary considerably based on the genetic variety of the cotton [13–15]. In addition, Pons et al. [16] reported that the gossypol content of cottonseeds was negatively correlated with the temperature and positively correlated with the rainfall that the cotton plants were exposed to while producing the seed. Furthermore, methods of cottonseed oil extraction that achieve maximum oil production yield CSM with the lowest gossypol content [1]. Finally, CSM that contains soapstock has a higher concentration of gossypol than a meal without added soapstock because soapstock, a waste by-product from the cottonseed oil extraction process, contains a high concentration of gossypol [17, 18].

For CSM to be utilized as a feed ingredient in poultry diets, a quick and reliable method for the determination of gossypol is needed. The official methods for the quantification of free and total gossypol in CSM are those of the AOCS [19, 20]. However, these procedures are time-consuming as well as laboratory- and labor-intensive. A quicker alternative method for determining the gossypol content of CSM might be near infrared reflectance spectroscopy (NIRS). As reviewed by Van Kempen [21], NIRS technology has been used for the past several years to quickly and efficiently determine moisture, protein, fat, metabolizable energy, total amino acid, and digestible amino acid contents in feed ingredients and feeds.

Previous attempts have been made to measure the gossypol content of whole cottonseeds by NIRS [22, 23]. There are no reports, however, of predicting the gossypol content of CSM by NIRS. Therefore, it is the goal of the present study to develop a NIRS calibration prediction equation to analyze the free gossypol content of CSM.


   MATERIALS AND METHODS
TOP
 SUMMARY
 DESCRIPTION OF PROBLEM
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS AND APPLICATIONS
 REFERENCES AND NOTES
 
Experiment 1
Ninety-five CSM samples were obtained from various locations around the world (Table 1Go). All of the CSM samples were ground through a 1-mm sieve in a Wiley Laboratory Intermediate Mill [24] to obtain a uniform particle composition across all samples. The free gossypol content of a portion of each CSM sample was then chemically determined by the AOCS method [19].


View this table:
[in this window]
[in a new window]

  		Table 1. Free gossypol content of cottonseed meal samples obtained from several countries
 
A portion of each CSM sample was also scanned for gossypol using a model 6500 monochromator NIRS system [25]. The NIRS system continuously scanned wavelengths from 1,100 to 2,500 nm. Reflectance data were recorded at 2-nm intervals. The wavelengths used for free gossypol calibration were based on the analysis of the pure gossypol sample [26] and previous reports of free gossypol determination in whole cottonseeds by NIRS [22, 23]. The NIRS spectra data were then processed using the software WinISI 11 [27]. The spectra data from 56 randomly selected samples of the original 95 samples were then combined with the AOCS reference data to generate a calibration prediction equation using a modified partial least-squares regression method [27].

Once the prediction equation was obtained it was validated with the remaining CSM samples. To validate the prediction equation, a linear regression analysis was used to compute the correlation between the gossypol content obtained for these samples by the AOCS method and the NIRS method. The validation procedure was used to estimate the r2 and standard error of prediction (SEP) as parameters for calibration accuracy. The validation was also used to determine the average difference between the gossypol reference method and predicted values (bias), and the linear regression of the gossypol reference method against predicted values (slope).

Experiment 2
Many of the international samples used in experiment 1 appeared to be poorly processed based on the presence of a large amount of residual cotton fibers and the very dark appearance of the meal. Furthermore, the oil extraction procedures for the international samples were not known. Therefore, in experiment 2 only the more uniformly processed CSM samples obtained from the United States were used for the NIRS calibration. Thus, a total of 51 samples from the original 95 were used in this experiment. Because the sample population was small, the validation sample set was composed of the same 51 samples used for the calibration. The CSM samples from the United States were obtained from 15 different oil processing mills distributed across 8 states (AR, CA, GA, MS, OK, SC, TN, and TX). When CSM samples were obtained from the same oil mill, the samples were collected from different batches of meal production.


   RESULTS AND DISCUSSION
TOP
 SUMMARY
 DESCRIPTION OF PROBLEM
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS AND APPLICATIONS
 REFERENCES AND NOTES
 
Experiment 1
As determined by the AOCS method, the free gossypol content of CSM samples had a wide distribution from 0.01 to 0.39% (Figure 1Go). Obtaining CSM samples with this high variability in free gossypol content was critical because a NIRS prediction equation should be constructed from samples that represent the range of values that will be encountered in field testing. The CSM samples obtained from Egypt and Peru contained a high concentration of gossypol, whereas the samples from Tajikistan and Iran had lower gossypol levels (Table 1Go).


Figure 1
View larger version (15K):
[in this window]
[in a new window]

  		Figure 1. Distribution of the free gossypol content determined by the official methods of the American Oil Chemists’ Society (AOCS) [19] in the cottonseed meal (CSM) samples analyzed from several countries.

 
The regression analysis for the validation of the reflectance values for gossypol in the international CSM samples yielded an r2 of 0.728 and SEP of 0.034 (Figure 2Go). Although this coefficient of determination is statistically acceptable, a higher correlation between the reference and reflectance values was desirable.


Figure 2
View larger version (14K):
[in this window]
[in a new window]

  		Figure 2. Correlation between the free gossypol content of the international cottonseed meal samples determined by the American Oil Chemists’ Society (AOCS) [19] official method and the predicted content based on near infrared reflectance (NIRS) measurements.

 
Experiment 2
Cottonseed oil extraction in the United States is very efficient and yields a uniform high quality CSM that is devoid of excess lint and residual oil from the cottonseed [1, 28]. Therefore, it was not surprising that when only the samples obtained from the United States were utilized for the NIRS prediction equation, the regression analysis of the gossypol determinations between the AOCS and the NIRS methods yielded a higher r2 of 0.921 and a SEP of 0.014 (Figure 3Go).


Figure 3
View larger version (13K):
[in this window]
[in a new window]

  		Figure 3. Correlation between the free gossypol content of the US cottonseed meal samples determined by the American Oil Chemists’ Society (AOCS) [19] official method and the predicted content based on near infrared reflectance (NIRS) measurements.

 
Because only 51 CSM samples from the United States were available, the same samples had to be used to formulate and validate the NIRS free gossypol prediction equation. Ideally, the validation would have been done as in experiment 1, with new CSM samples. For comparison purposes, we also determined that if the validation sample set consisted of the same 56 CSM samples used to formulate the prediction equation for the international samples (experiment 1), the r2 value actually decreased to 0.652 and the SEP value remained at 0.034. Therefore, we feel confident that the NIRS free gossypol prediction equation that was constructed in experiment 2 can be utilized to accurately determine the free gossypol content of CSM produced in the United States.

To obtain an accurate free gossypol concentration in CSM samples using NIRS, the results from the 2 current experiments suggest that NIRS prediction equations for free gossypol content of CSM may have to be constructed for different geographical regions of the world. The need for different free gossypol prediction equations reflects the differences in CSM oil extraction processing in different parts of the world, which result in CSM that varies tremendously in quality and appearance. Many of the international CSM samples contained an excessive amount of residual lint compared with the CSM produced in the United States. This lint may interfere with accurate free gossypol determinations by NIRS. Wadsworth and Richard [23] reported a very high SEP when they developed an NIRS calibration for gossypol in whole fuzzy (covered in lint) cottonseed.

The main purpose of constructing a NIRS prediction equation in 1 NIRS system is to then transfer that prediction equation to other NIRS instruments. However, no 2 NIRS systems are the same, and even subtle differences between them may cause variation in the spectral data. Such differences can make the transfer of calibration equations ineffective. Fortunately, there are mathematical manipulations that use the slope and bias factors obtained when each prediction equation is constructed, which can be used to correct spectral data between NIRS analyzers. For the international samples, the slope and bias values were 1.034 and –0.003, respectively, whereas for the US samples the slope and bias were 1.0 and 0.0, respectively. Additionally, instead of using the bias and the slope correction method, a new calibration equation for free gossypol could also be quickly established for a different NIRS system by utilizing a preestablished CSM sample set in which the reference (AOCS) was already determined.


   CONCLUSIONS AND APPLICATIONS
TOP
 SUMMARY
 DESCRIPTION OF PROBLEM
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS AND APPLICATIONS
REFERENCES AND NOTES
 

  1. A NIRS calibration was successfully developed for the determination of free gossypol in CSM.
  2. Due to differences in CSM processing around the world, separate prediction equations for individual regions of the world may be needed to confidently predict the free gossypol content of CSM by NIRS.
  3. Cottonseed processing plants in the United States could use NIRS as a rapid means of quantifying the gossypol content of the CSM they produce. Once the gossypol content was determined, poultry nutritionists and feed mill operators could decide whether to purchase the CSM for use in poultry diets.


   FOOTNOTES
 
1 This research was supported in part by grant 05-635GA from the Georgia Cotton Commission, Perry, GA. Back


   REFERENCES AND NOTES
TOP
 SUMMARY
 DESCRIPTION OF PROBLEM
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS AND APPLICATIONS
 REFERENCES AND NOTES
 

  1. Waldroup, P. W. 1981. Cottonseed meal in poultry diets. Feedstuffs 53:21–24.
  2. Heywang, B. W., H. R. Bird, and R. P. Kupperman. 1952. The loss or inactivation of pure gossypol in a mixed diet. Poult. Sci. 31:35–39.[Web of Science]
  3. Phelps, R. A. 1966. Cottonseed meal for poultry: From research to practical application. World’s Poult. Sci. J. 22:86–112.[CrossRef]
  4. Lillie, R. J., and H. R. Bird. 1950. Effect of oral administration of pure gossypol and of pigment glands of cottonseed on mortality and growth of chicks. Poult. Sci. 29:390–393.[Web of Science]
  5. Milligan, J. L., and H. R. Bird. 1951. Effect of processing variants on the nutritive value of cottonseed meal for chicks. Poult. Sci. 30:651–657.[Web of Science]
  6. Lordelo, M. M., A. J. Davis, M. C. Calhoun, M. K. Dowd, and N. M. Dale. 2005. Relative toxicity of gossypol enantiomers in broilers. Poult. Sci. 84:1376–1382.[Abstract/Free Full Text]
  7. Couch, J. R., W. Y. Chang, and C. M. Lyman. 1955. The effect of free gossypol on chick growth. Poult. Sci. 34:178–183.[Web of Science]
  8. Heywang, B. W., and H. R. Bird. 1955. Relationship between the weight of chicks and levels of dietary free gossypol supplied by different cottonseed products. Poult. Sci. 34:1239–1247.[Web of Science]
  9. Henry, M. H., G. M. Pesti, R. Bakalli, J. Lee, R. T. Toledo, R. R. Eitenmiller, and R. D. Phillips. 2001. The performance of broiler chicks fed diets containing extruded cottonseed meal supplemented with lysine. Poult. Sci. 80:762–768.[Abstract/Free Full Text]
  10. Waldroup, P. W., E. G. Keyser, V. E. Tollett, and T. E. Bowen. 1968. The evaluation of a low-gossypol glandless cottonseed meal in broiler diets. Poult. Sci. 47:1179–1186.[Web of Science]
  11. Watkins, S. E., J. T. Skinner, M. H. Adams, and P. W. Waldroup. 1994. An evaluation of low-gossypol cottonseed meal in diets for broiler chickens 2. Influence of assigned metabolizable energy values and supplementation with essential amino acids on performance. J. Appl. Poult. Res. 3:7–16.[Abstract/Free Full Text]
  12. Lordelo, M. M., A. J. Davis, J. L. Wilson, and N. M. Dale. 2004. Cottonseed meal diets improve body weight uniformity in broiler breeder pullets. J. Appl. Poult. Res. 13:191–199.[Abstract/Free Full Text]
  13. Boatner, C. H., L. E. Castillon, C. M. Hall, and J. W. Neely. 1949. Gossypol and gossypurpurin in cottonseed of different varieties of G. barbadense and G. hirsutum, and variation of the pigments during storage of the seed. J. Am. Oil Chem. Soc. 26:19–25.[CrossRef][Web of Science]
  14. Sullivan, J. L., J. T. Huber, and J. M. Harper. 1993. Performance of dairy cows fed short staple, Pima, and cracked Pima cottonseed and feed characteristics. J. Dairy Sci. 76:3555–3561.[Abstract/Free Full Text]
  15. Santos, J. E., P. M. Villasenor, E. J. DePeters, P. H. Robinson, and B. C. Baldwin. 2002. Type of cottonseed and gossypol in diets of lactating dairy cows: Lactation performance and plasma gossypol. J. Dairy Sci. 85:1491–1501.[Abstract]
  16. Pons, W. A., Jr., C. L. Hoffpauir, and T. H. Hopper. 1953. Gossypol in cottonseed—Influence of variety and environment. J. Agric. Food Chem. 1:1115–1118.[CrossRef][Web of Science]
  17. Dowd, M. K. 1996. Compositional characterization of cottonseed soapstocks. J. Am. Oil Chem. Soc. 73:1287–1295.[CrossRef][Web of Science]
  18. Davis, A. J., M. M. Lordelo, and N. M. Dale. 2002. The use of cottonseed meal with or without added soapstock in laying hen diets. J. Appl. Poult. Res. 11:127–133.[Abstract/Free Full Text]
  19. American Oil Chemists’ Society. 1985. Official Methods Ba 7-58: Determination of free gossypol. Official and Tentative Methods of Analysis. 3rd ed. Am. Oil Chem. Soc., Chicago, IL.
  20. American Oil Chemists’ Society. 1985. Official Methods Ba 8-78: Determination of total gossypol. Official and Tentative Methods of Analysis. 3rd ed. Am. Oil Chem. Soc., Chicago, IL.
  21. VanKempen, T. 2001. Infrared technology in animal production. World’s Poult. Sci. J. 57:29–48.[CrossRef][Web of Science]
  22. Birth, G. S., and H. H. Ramey. 1982. Near-infrared reflectance for analysis of cottonseed for gossypol. Cereal Chem. 59:516–519.
  23. Wadsworth, J. I., and O. L. Richard. 1993. Analysis of fuzzy cottonseed by NIR: Calibration update. Pages 69–92 in Proc. 42nd Oilseed Processing Clinic, New Orleans, LA. USDA-ARS, New Orleans, LA.
  24. Thomas Scientific, Swedesboro, NJ.
  25. Foss North America, Eden Prairie, MN.
  26. Sigma Chemical Company, St. Louis, MO.
  27. WinISI. 2000. The complete Software Solution for Routine Analysis and Robust Calibrations. Version 1.50. Foss NIRSystems Inc., Silver Springs, MD.
  28. Jones, L. A. 1981. Nutritional values for cottonseed meal. Feedstuffs 53:19–21.




This Article
Right arrow Summary Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lordelo, M. M.
Right arrow Articles by Davis, A. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Lordelo, M. M.
Right arrow Articles by Davis, A. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS