Department of Animal Science, University of Padova, Viale dell'Università 16, 35020 Legnaro (PD), Italy.
J Anim Sci. 2012 Feb;90(2):429-38. doi: 10.2527/jas.2011-4150. Epub 2011 Sep 23.
The aims of this study were 1) to investigate the potential application of near-infrared spectroscopy (NIRS) to predict intramuscular fat (IMF) and fatty acid (FA) composition of individual meat samples, 2) to estimate heritability of IMF and FA NIRS-based predictions, and 3) to assess the statistical relevance of the genetic background of such predictions by using the Bayes factor (BF) procedure. Young Piemontese bulls (n = 1,298) were raised and fattened on 124 farms, and slaughtered at the same commercial abattoir. Intramuscular fat content and FA composition were analyzed on a random subset of 148 samples of minced and homogenized longissimus thoracis muscle. Near-infrared spectroscopy spectra were collected on all samples (n = 1,298) in reflectance mode between 1,100 and 2,498 nm (every 2 nm) using fresh minced meat samples. Calibration models developed from the random subset of 148 samples were used to predict IMF and FA contents of the remaining 1,150 samples. Intramuscular fat content and FA predictions were analyzed under a Bayesian univariate animal linear models, and the statistical relevance of heritability estimates was assessed through BF; the model with polygenic additive effects was favored when BF > 1. In general, satisfactory results (R(2) > 0.60) were obtained for 6 out of the 8 major FA (C14:0, C:16:0, C16:1, C18:0, C18:1n-9 cis/trans, and C18:1n-11 trans), 6 out of the 19 minor FA (C10:0, C12:0, C17:0, C17:1, C18:2 cis-9,trans-11, and C20:2), and the total SFA, MUFA, and PUFA. Bayes factors between models with and without a genetic component provided values greater than 1 for IMF, C14:0, C16:0, C18:1n-9 cis/trans, C17:0, C17:1, C20:2, SFA, MUFA, and PUFA. The greatest BF was reached by C20:2 (BF >10), suggesting strong evidence of genetic determinism, whereas IMF, C18:1n-9 cis/trans, C17:0, C17:1, MUFA, and PUFA showed substantial evidence favoring the numerator model (3.16 < BF < 10). Point estimates of heritabilities for FA predicted by NIRS were low to moderate (0.07 to 0.21). Results support that NIRS is a useful technique to satisfactorily predict some FA of meat. The existence of an important genetic determinism affecting FA profile has been confirmed, suggesting that FA composition of meat can be genetically modified.
1)探讨近红外光谱(NIRS)在预测单个肉样的肌内脂肪(IMF)和脂肪酸(FA)组成方面的潜在应用;2)估计 IMF 和 FA 的 NIRS 预测的遗传力;3)通过贝叶斯因子(BF)程序评估遗传背景对这些预测的统计学相关性。124 个农场饲养了年轻的皮埃蒙特公牛(n=1298),并在同一商业屠宰场育肥。随机抽取 148 份切碎和匀浆的背最长肌样本,分析肌内脂肪含量和 FA 组成。使用新鲜的肉末样本,在反射模式下,在 1100nm 至 2498nm(每隔 2nm)范围内采集所有 1298 个样本的近红外光谱。从随机选择的 148 个样本中开发的校准模型用于预测其余 1150 个样本的 IMF 和 FA 含量。采用单变量动物线性模型下的贝叶斯分析了肌内脂肪含量和 FA 预测值,并通过 BF 评估了遗传力估计的统计学相关性;当 BF>1 时,优先考虑具有多基因加性效应的模型。一般来说,对于 8 种主要 FA(C14:0、C:16:0、C16:1、C18:0、C18:1n-9 cis/trans 和 C18:1n-11 trans)中的 6 种、19 种次要 FA(C10:0、C12:0、C17:0、C17:1、C18:2 cis-9,trans-11 和 C20:2)以及总 SFA、MUFA 和 PUFA,获得了令人满意的结果(R2>0.60)。IMF、C14:0、C16:0、C18:1n-9 cis/trans、C17:0、C17:1、C20:2、SFA、MUFA 和 PUFA 的模型间和模型内的贝叶斯因子均大于 1。C20:2 的 BF 值最大(BF>10),表明遗传决定作用具有很强的证据,而 IMF、C18:1n-9 cis/trans、C17:0、C17:1、MUFA 和 PUFA 则显示出明显的优势,支持分子模型(3.16<BF<10)。NIRS 预测 FA 的遗传力估计值较低至中等(0.07 至 0.21)。结果支持 NIRS 是一种非常有用的技术,可以令人满意地预测一些肉类的 FA。FA 图谱的重要遗传决定作用已经得到证实,这表明可以对肉的 FA 组成进行遗传修饰。