The Use of Noninvasive Measurements for Predicting Objective Tenderness of Muscles from the Beef Rou

1
The Use of Non-invasive Measurements for
Predicting Objective Tenderness of Muscles from
the Beef Round

• J.T. Sawyer, J.K. Apple, J.-F. Meullenet, B.
  Cheatman, W.K. Chung, R. Xiong, and S. G. Bajwa

2
Introduction

• Numerous factors affect the quality of meat
  (Koohmaraie, 1995)
• Breed, age, sex, weight, environment
• 70 of variation in beef tenderness explained by
  the environment and non-additive gene effects
• Extreme difficulty for consumer selections
• Meat quality has become perceptually dependent
  upon three criteria
• Tenderness, color, and shelf-stability
• Consumer acceptance highly dependent upon
  tenderness (Jeremiah, 1982 Warkup, 1995)
• Implementation of quality guarantees through
  certified programs
• Assessment of meat quality parameters
• Time consuming
• Destructive
• Expensive

3
Historical Perspective

• Liu, 2003
• WBSF
• Reflectance 400 to 2,498 nm
• Aging (days)
• 2 R2 0.41 to 0.48
• 4 R2 0.64 to 0.69
• 8 R2 0.22 to 0.18
• 14 R2 0.17 to 0.72
• 21 R2 0.24 to 0.56
• Sensory
• Chewiness R2 0.38 to 0.58
• Juiciness R2 0.18 to 0.50

• Tremendous amount of variation in tenderness
  prediction
• Mitsumoto, 1991
• WBSF
• Reflectance R2 0.62
• Transmittance R2 0.68
• Fiber optic R2 0.68
• Bryne, 1998
• WBSF
• 750 to 1,180 nm
• R2 0.61 to 0.82
• Leroy, 2003
• WBSF
• Reflectance R2 0.25 day 2 0.19 day 8
• Transmittance R2 0.41 day 2 0.15 day 8

4
Near Infrared

• Near infrared (NIR) is part of the infrared light
• Has longer wavelength than visible light
• from the end of visible light 400nm to 2,500nm
• Near Infrared (NIR) spectroscopy has gained
  increased interest, especially in process control
  applications

5
How does it work?

• NIR spectroscopy measuring absorbance or
  reflectance of different NIR frequencies by a
  sample
• Different functional groups absorb characteristic
  frequencies
• The main goal of NIR spectroscopic analysis is to
  determine the chemical functional groups in a
  sample

6
Instruments

• NIRSystems 6500 (Foss) outfitted with a fiber
  optic probe
• Reflectance monochrometer in the range of
  400-2,400nm
• Analytical Spectral Device (Field Spec Pro)
• Reflectance in the range of 350-1,050 nm

7
Instrumental Applications

• NIR on-line analysis (Osborne, 1986)
• Remote non-contact
• Low cost
• Susceptible to ambient light
• Dust build-up
• Atmospheric humidity
• By-Pass sampler
• Most popular
• Stationary
• Fiber Optic probe
• Widest range of application
• Data collection
• 780 to 2,500 nm

• Food product applications
• Cereal
• Dairy
• Meat
• Fish
• Fruit vegetables
• Confectionary
• Beverages
• Authenticity

8
Objective

• Comparison of two spectroradiometers (NIRS vs.
  ASD) and instrumental tenderness measures (WBSF
  vs. RBSF), and develop an empirical model that
  could accurately predict tenderness across
  diverse aging periods.

9
Materials and Methods

• Beef top (inside) and bottom rounds (gooseneck
  IMPS 168 170) randomly collected from four
  USDA quality grades (n10/QG) during fabrication.
• Prime, Top Choice (CAB), Low Choice, Select
• Steaks (n5/subprimal) were allotted to one of
  five aging periods
• Aging periods (0, 7, 14, 21, 28)

10
Spectroradiometer Acquisition

• All steaks (n600) were initially scanned
• NIRSystem spectroradiometer
• 400 to 2,400 nm
• ASD Field Spec Pro portable spectroradiometer
• 350 to 1,050 nm
• Vacuum-packaged and aged in the absence of light
  to one of the five aforementioned aging periods
• Steaks were rescanned upon completion of each
  designated aging period
• Frozen at -20º C until measured for instrumental
  tenderness

11
Instrumental Tenderness Assessment

• Frozen steaks thawed for 16 h at 2 ºC
• Cooked in a commercial convection oven preheated
  to 165 ºC
• Meullenet-Owens shear force (MORS)
• Six measurements
• 4 scanned locations 2 random locations/steak
• 4.99-kg load cell
• Blade height 24 mm blade width 8 mm
• 10 mm/sec travel speed
• 20 mm penetration depth
• Texture Analyzer (Texture Technologies,
  Scarsdale, NY)
• Warner-Bratzler shear force (WBSF)
• Six cores
• 4 cores scanned location
• 55-kg tension/compression load cell
• 250 mm/min crosshead speed
• Universal Testing Machine (Instron Corp., Canton,
  MA)

12
Statistical Modeling

• Inverse raw reflectance data were averaged and
  processed in 2-nm intervals
• Second derivative of the spectra obtained on a
  20-point scale
• Savitzky Golay algorithm (Esbenson, 2002)
• Raw data plotted with normal probability plot
  option to ensure normality
• Instrumental tenderness data
• PLS regression included
• PLS1 option (CAMO, Oslo, Norway)
• Entire spectral range
• NIR 400 to 2,400 nm
• ASD 350 to 1,050 nm
• Spectral data centered

13
Model Validation

• Ensures reliability of the model for future use
• Full cross-validation recommended procedure by
  CAMO
• Jack-knife option used
• Select significant wavelengths
• Additional analysis
• PROC GLM (SAS Inst. Inc., Cary, NC)
• Least Squares Means separated with PDIFF option

14
Instrumental Tenderness
15
Instrumental Tenderness
16
Instrumental Tenderness
17
Instrumental Tenderness
18
Instrumental Tenderness
19
Instrumental Tenderness
20
Biceps Femoris Spectral
21
Biceps Femoris Spectral
22
Semitendinosus Spectral
23
Semitendinosus Spectral
24
Semimembranosus Spectral
25
Semimembranosus Spectral
26
Implications

• NIRS was a better method for tenderness
  prediction
• ASD produced promising results for MORS
• NIRS successful in predicting WBSF and MORS for
  the semimembranosus
• Instrumental methods are not comparable due to
  individual patterns of prediction and behavior
• Future studies are essential to assist in
  scalability, predictive ability, and precision

27

• QUESTIONS