Pharmacometabolomics with a combination of PLS‐DA and random forest algorithm analyses reveal meloxicam alters feline plasma metabolite profiles

Abstract

Repeated administration of meloxicam to cats is often limited by the potential damage to multiple organ systems. Identifying molecules that predict the adverse effects of meloxicam would help to monitor and individualize its administration, maximizing meloxicam's beneficial effects. The objectives of this study were to (a) determine if the repeated administration of meloxicam to cats alters the plasma metabolome and (b) identify plasma metabolites that may serve to monitor during the administration of meloxicam in cats. Purpose bred young adult cats (n = 12) were treated with meloxicam at 0.3 mg/kg or saline subcutaneously once daily for up to 17 days. An untargeted metabolomics approach was applied to plasma samples collected prior to and at designated time points after meloxicam or saline administration. To refine the discovery of biomarkers, the machine‐learning algorithms, partial least squares discriminant analysis (PLS‐DA) and random forest (RF), were trained and validated using a separate unrelated group of meloxicam‐ and saline‐treated cats (n = 8). A total of 74 metabolites were included in the statistical analysis. Metabolomic analysis shows that the repeated administration of meloxicam alters multiple substances in plasma, including nonvolatile organic acids, aromatic amino acids, monosaccharides, and inorganic compounds as early as four days following administration of meloxicam. Seventeen plasma molecules were able to distinguish meloxicam‐treated from saline‐treated cats. The metabolomic changes discovered in this study may help to unveil unknown mechanisms of NSAID‐induced side effects. In addition, some metabolites could be valuable for individualizing the administration of meloxicam to cats to mitigate adverse effects.

UHPLC–MS/MS-Based Nontargeted Metabolomics Analysis Reveals Biomarkers Related to the Freshness of Chilled Chicken

“…As a machine learning method, the random forest model is a regression tree technique that uses bootstrap aggregation and randomization of predictors to achieve a high degree of predictive accuracy [ 27 ]. The random forest model has been widely used to analyze complex metabolomics data [ 28 , 29 , 30 , 31 , 32 , 33 ] and has its unique advantages. First, it is relatively robust to outliers and noise.…”

Metabolomics meets systems immunology

“…The chemometrics analysis methods such as principal component analysis (PCA), partial least squares discriminant analysis (PLS‐DA), orthogonal partial least squares discriminant analysis (orthoPLS‐DA), and sparse partial least squares discriminant analysis (sPLS‐DA) are also widely used (Rio et al , 2009 ; Miolo et al , 2016 ; He et al , 2018 ). Clustering analysis approaches including hierarchical clustering, K‐means, and self‐organizing map (SOM), as well as classification and feature selection such as random forest (RF) and support vector machine (SVM), are commonly applied to metabolomics studies (Bartel et al , 2013 ; Li et al , 2016 ; Broughton‐Neiswanger et al , 2020 ). MetaboAnalyst also offers biological interpretation tools such as pathway analysis, enrichment analysis, and network analysis.…”

The Serum and Fecal Metabolomic Profiles of Growing Kittens Treated with Amoxicillin/Clavulanic Acid or Doxycycline

“…Previous studies have focused on the taxonomical and compositional characterization of the GI microbiome in young cats [11][12][13][14][15][16][17][18]. More recent studies have also described the serum and fecal metabolites in cats in states of health [19][20][21][22][23] and disease [24][25][26][27][28][29], or following drug administration [30][31][32][33]. Investigating metabolic patterns under certain conditions has filled gaps in understanding cellular processes and has led to the discovery of new disease biomarkers, allowing an understanding of impaired signaling pathways in different disease states [9].…”