Rune Rødbotten - abstract fra doktorgradsavhandling

Ane Gro Siri Skjelfjord

Abstract fra avhandlingen

Prediction of beef tenderness by NIR spectroscopy, and examination of the influence of ante- and post-mortem factors on tenderness / Prediksjon av mørhet i storfekjøtt ved bruk av nær-infrarød spektroskopi, og undersøkelse av ante- og post-mortem faktorers påvirkning på mørhet

The theme of this thesis is tenderness in bovine longissimus dorsi. Today, there is large variability in tenderness of beef offered for purchase, which is a problem for consumers and thus also the meat industry. The present work has been focused in two different directions. One of these was examination of the influence of some ante- and post-mortem factors on final tenderness. The other direction was prediction of tenderness from near infrared (NIR) spectra.

Angus x NRF and Limousin X NRF bulls were fed two different levels of concentrate in a controlled experiment. It was found that concentrate level affected growth rate of the animals, but neither feeding level nor crossbreed was significant for final tenderness of the longissimus dorsi. Three post mortem factors were studied. Chilling rate of the muscles in the first 24 hours after slaughter and ageing time of the meat samples was found highly significant for tenderness of the beef samples. Low voltage electrical stimulation (90V, 15Hz, 20 sec) was applied within 10 minutes post mortem, but no significant effect on tenderness was obtained for this treatment. However, there was an interaction between electrical stimulation and chilling rate. For the rapidly chilled beef samples there was a tendency of increased tenderness when electrical stimulation was applied, while there was a tendency in opposite direction at moderate chilling conditions.

Two different strategies were studied for prediction of tenderness based on NIR spectra of the beef samples. Tenderness was either predicted at the same time as the spectra were recorded, whereas the other approach was to do forward prediction about tenderness level after the beef samples had been aged for a period. Multivariate correlation coefficients in the range 0.60 – 0.85 were obtained when tenderness was predicted at the same time as spectra was recorded, while the corresponding range was 0.47 – 0.55 when forward prediction was performed. For the majority of models reported the highest correlation coefficients were achieved when multiplicative scatter correction (MSC) treatment was applied to the NIR spectra prior to modelling. Although fair correlation coefficients were obtained for some of the multivariate prediction models, prediction errors (RMSEP) of the models should be decreased before NIR spectroscopy could be recommended for commercial prediction of tenderness. Although some samples were misclassified in this study it seems like a classification approach could be used to segregate between the most and least tender beef samples.

Pr. desember 2004:
Rune Rødbotten, Borregaard, Forskningsbygget, Postboks 162, 1701
Sarpsborg

Publisert: 01.12.08
Oppdatert: 05.01.09
Utskriftsvennlig versjon

Del med en venn:

Institutt for husdyr- og akvakulturvitenskap

Tlf: +47 64 96 51 00
E-post:iha@umb.no

Webansvarlig: Janne Karin Brodin

Emneord

- Forskning
- Husdyrfag
- Person
- Storfe
- Utdanning