Search

Our objective was to validate the possibility of detecting SARA from milk Fourier transform mid-infrared spectroscopy estimated fatty acids (FA) and machine learning. Subacute ruminal acidosis is a common condition in modern commercial dairy herds for which diagnosis remains challenging due to its symptoms often being subtle, nonexclusive, and not immediately apparent. This observational study aimed at evaluating the possibility of predicting SARA by developing machine learning models to be applied to farm data and to provide an estimated portrait of SARA prevalence in commercial dairy herds. A first dataset, composed of 488 milk samples from 67 cows (initial DIM = 8.5 ± 6.18; mean ± SD) from 7 commercial dairy farms and their corresponding SARA classification (SARA+ if rumen pH <6.0 for 300 min, otherwise SARA-) was used for the development of machine learning models. Three sets of predictive variables (milk major components [MMC], milk FA [MFA], and MMC combined with MFA [MMCFA]) were submitted to 3 different algorithms, namely elastic net (EN), extreme gradient boosting, and partial least squares, and evaluated using 3 different scenarios of cross-validation. The accuracy, sensitivity, and specificity of the resulting 27 models were analyzed using a linear mixed model. Model performance was not significantly affected by the choice of algorithm. Model performance was improved by including FA estimations (MFA and MMCFA as opposed to MMC alone). Based on these results, 1 model was selected (algorithm: EN; predictive variables: MMCFA; 60.4%, 65.4%, and 55.3% of accuracy, sensitivity, and specificity, respectively) and applied to a large dataset comprising the first test-day record (milk major components and FA within the first 70 DIM of 211,972 Holstein cows [219,503 samples]) collected from 3,001 commercial dairy herds. Based on this analysis, the within-herd SARA prevalence of commercial farms was estimated at 6.6 ± 5.29% ranging from 0% to 38.3%. A subsequent linear mixed model was built to investigate the herd-level factors associated with higher within-herd SARA prevalence. Milking system, proportion of primiparous cows, herd size, and seasons were all herd-level factors affecting SARA prevalence. Furthermore, milk production was positively associated with SARA prevalence, and milk fat yield was negatively associated. Due to their moderate levels of accuracy, the SARA prediction models developed in our study, using data from continuous pH measurements on commercial farms, are not suitable for diagnostic purposes. However, these models can provide valuable information at the herd level., (The Authors. Published by Elsevier Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).)

A 5-year strategic research network with a diverse base of industry, government, and academic partners was approved for support by National Sciences and Engineering Research Council of Canada (NSERC) on January 3, 2005. This Metals in the Human Environment Strategic Network (MITHE-SN) builds on, and further extends, science knowledge developed by the NSERC-sponsored Metals in the Environment Research Network (MITE-RN, 1999-2004). In addition to the initial award, the MITHE-SN received an additional 2-year grant specifically targeted to (1) enhance training opportunities for internships with international organizations, (2) increase international networking and linkages, and (3) optimize knowledge dissemination and technology transfer. The research program is comprised of three themes and represents a cascade of effects along food webs, from the lowest trophic levels to the highest consumers. Each of the themes addresses issues related to distinguishing the magnitudes and roles of natural background and anthropogenic metal inputs in biotic exposure to metals; estimating the bioavailable fraction of metals in the exposure media, thus better quantifying the true exposure concentration; and determining the factors that influence bioavailability of metals in media, so that predictive models can be developed for use in the development of site-specific metals criteria.