PubAg

Main content area

Proposal and validation of new indexes to evaluate maize silage fermentative quality in lab-scale ensiling conditions through the use of a receiver operating characteristic analysis

Author:
Andrighetto, Igino, Serva, Lorenzo, Gazziero, Matteo, Tenti, Sandro, Mirisola, Massimo, Garbin, Elisabetta, Contiero, Barbara, Grandis, Daniel, Marchesini, Giorgio
Source:
Animal feed science and technology 2018 v.242 pp. 31-40
ISSN:
0377-8401
Subject:
ammonia, analysis of variance, butyric acid, corn, corn silage, cow feeding, dairy cows, data collection, ethanol, fermentation, lactic acid, near-infrared spectroscopy, pH, silage making, society
Abstract:
In the context of dairy cow feeding, it is increasingly important to know the quality of the maize silage used in the ration and therefore, it appears to be crucial optimizing the techniques necessary to assess it. The aim of this study was to evaluate whether some reference indexes, like the Flieg-Zimmer’s (FZS), the German agricultural society’s (DLG) and Vanbelle’s scores, could properly estimate the quality of fermentations of maize silage made in a lab-scale ensiling system, and to calculate and validate new quality indexes suitable for lab-scale fermentations. The experimental dataset was obtained by analysing through near-infrared spectroscopy 522 samples of whole maize crop ensiled immediately after the harvest, using the vacuum-packing technique. The six (I1 – I6) new indexes were calculated on the basis of chemical and physical parameters as: pH, organic acids, ethanol, etc. All the indexes were tested for normality with the Shapiro–Wilk test. In order to define the accuracy with which the new indexes ranked the maize silage on the basis of its fermentation quality, a receiver operating characteristic (ROC) analysis was performed, using the FZS as gold standard test and dichotomizing the FZS in two levels according to a cut-off (FZS < 80, non-excellent vs. FZS ≥ 80, excellent). Accuracy was determined through the value of the area under the curve (AUC). Finally, a one-way ANOVA was used to compare the quality of maize silage with low (< 320 g/kg), medium (320–360 g/kg) and high (> 360 g/kg) dry matter (DM). In the lab-scale silages the new indexes were normally distributed, whereas the reference indexes were not. The new indexes showed values of AUC ranging between 0.76 and 0.89, with the I5 index showing the best combination of sensitivity (0.87) and specificity (0.77) in discriminating between good and poor quality silage. The cut-off of the new indexes ranged between 45.0 and 57.4 points. The lab-scale silages were all excellent, no matter the category of DM. However, while FZS and DLG did not differ among the 3 categories, I1 – I6 were significantly higher in silages with low DM (P < 0.001). Silages with low DM had the highest concentrations of lactic acid (56.4 g/kg DM, P < 0.001), ammonia (61.4 g/kg DM, P < 0.001) and butyric acid (0.62 g/kg DM, P < 0.001) as well. Data confirmed that the new proposed indexes are promising in describing the fermentation quality of maize silage in lab-scale conditions.
Agid:
6053827