Introduction
Poultry is an important product in the global meat market. While most foods of animal origin are excellent sources of high-quality protein, many are also high in fat. The amount and quality of fat in all meat products is a key ingredient and poultry is one alternative for health-conscious consumers who want a lower fat content in their meat consumption. Quality assurance of both whole and minced chicken meat is critical in the poultry industry and NIR spectroscopy has been studied as an analytical tool for this. Fat, protein, and moisture are the main constituents of interest and the potential for measuring these parameters using NIR spectroscopy has been tested in various studies. Fatty acid profiles have also been studied. Even though chicken breast muscle has low-fat content, studies show good correlation for measuring fat and fatty acid profiles in ground chicken but there are particular challenges in measuring these parameters in whole chicken breast.
Analytes
- Fat
- Protein
- Moisture
- Fatty Acids Profile
Summary of Published Papers, Articles, and Reference Materials
NIR spectroscopy is an accurate and validated method for measuring fat, protein, moisture and fatty acids in many food products including poultry. However, the accuracy of calibration model predictions often varies for ground, homogenous chicken samples and whole chicken breasts. Comparative studies have shown worse results for whole breasts. Two possible reasons for this are the low-fat concentration in chicken breast muscle and the heterogeneity of the samples. Exceptional results have been shown for measuring some fatty acids in ground chicken meat as well as fat in chicken hamburgers. One promising study using whole chicken fillets in an online setting detected and graded wooden breast (WB) myopathy syndrome in chicken breast fillets. WB is a term for abnormal muscle tissue in the chicken breast, resulting in an unpleasant appearance. Chicken with WB is typically used to manufacture cheaper products, resulting in a loss to the producer. WB tissue has significantly higher moisture and less protein than healthy chicken breasts. Results showed that with proper calibration, detection of moisture and protein on intact breasts was good enough to sort WB and healthy breasts online. Multi-point quality measurement has potential as a Process Analytical Technology (PAT) tool for providing real-time feedback for process control.
Scientific References and Statistics
A Review of the Principles and Applications of Near-Infrared Spectroscopy to Characterize Meat, Fat, and Meat Products – Prieto, Pawluczyk, Dugan, Aalhus, Applied Spectroscopy. 2017, Vol. 71 (7) 1402-1426
Ground Chicken – Study 1
Ground chicken samples were scanned in transmittance over a short wavelength range (850 nm to 1050 nm). Results were expressed in absolute concentrations (% of total FA/ mg FA per 100 g-1 meat).
| Saturated Fatty Acids (SFA) | R2= 0.92 |
| Monounsaturated Fatty Acids (MUFA) | R2= 0.98 |
| Polyunsaturated Fatty Acids (PUFA) | R2= 0.65 |
| Omega-3 | R2= 0.39 |
| Omega-6 | R2= 0.67 |
Results are exceptional using transmission to predict SFA and MUFA and good enough for screening purposes to measure PUFA and Omega-6. The low concentrations of some individual PUFA was a likely factor in poor prediction as well as the greater number of double bonds in PUFA.
Freeze-Dried Ground Chicken – Study 2
Freeze-dried ground chicken samples were scanned in reflectance over a full VIS-NIR wavelength range (400 nm to 2500 nm). Results were expressed in absolute concentrations (% of total FA/ mg FA per kg-1 meat)
| Saturated Fatty Acids (SFA) | R2= 0.95 |
| Monounsaturated Fatty Acids (MUFA) | R2= 0.94 |
| Polyunsaturated Fatty Acids (PUFA) | R2= 0.97 |
| Omega-3 | R2= 0.95 |
| Omega-6 | R2= 0.98 |
Results were exceptional for all fatty acids in the freeze-dried samples. The freeze-drying process avoids water interference and increases FA concentrations, but it is costly and time-consuming as well as only being suited for off-line laboratory measurements.
http://journals.sagepub.com/doi/pdf/10.1177/0003702817709299
Fat in Chicken Hamburgers
Determination of Fat Content in Chicken Hamburgers Using NIR Spectroscopy and the Successive Projections Algorithm for Interval Selection in PLS Regression (iSPA-PLS) – Krepper, Romeo, Fernandes, et al., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, Volume 189, 15 January 2018, 300-306
Seventy chicken hamburger samples were scanned in reflectance mode ranging from 12.27 mg kg-1 to 32.12 mg kg-1. Various pre-processing of the spectral data were applied and tested in calibration models.
| Fat | R2= 0.94 |
Best results were shown by applying the Successive Projections Algorithm for interval selection in a Partial Least Squares regression. The method was successfully applied to chicken hamburger analysis and the results agreed with reference values at a 95% confidence level.
https://www.sciencedirect.com/science/article/pii/S1386142517306753
Wooden Breast (WB) Syndrome in Chicken Fillets
Rapid On-Line Detection and Grading of Wooden Breast Myopathy in Chicken Fillets by Near-Infrared Spectroscopy – Wold, Veiseth-Kent, Host, Lovland, PLOS ONE, DOI:10.1371, March 9, 2017
An industrial NIR scanner was chosen as a detection method for wooden breast (WB) syndrome in chicken breast fillets with a goal of using it for large scale on-line detection of the syndrome. One hundred ninety-seven fillets were used for the calibration set and seventy-nine fillets were used for a test set. The test set spectra were acquired under industry conditions to test the on-line feasibility of the system.
Two approaches were taken for classifying WB and healthy chicken fillets: Linear discriminant classification analysis using just the NIR spectra and a regression model using protein reference values as WB fillets have reduced protein concentration.
| Protein | R2= 0.76 |
| Moisture | R2= 0.67 |
While the correlation coefficients are low for both protein and moisture, this is understandable because of the relatively small range of values used (20.5% to 25.3% for protein). Prediction errors for both constituents were less than 0.6% and these results were good enough to separate WB and healthy fillets. Linear discrimination analysis of the test set revealed an optimum separation value of 21.9% protein between WB and healthy fillets. The calibration set obtained 99.5% correct classification and the test set scanned under industrial conditions obtained 100% correct classification, proving the feasibility of using the NIR scanner to classify WB and healthy chicken breast fillets.