In this work multivariate strategies were employed in order to highlight new potential biomarkers of interest to detect the exogenous treatment of steers intramuscularly treated with boldenone undecylenate. Serum samples collected from treated (n = 4) and control (n = 8) crossbred animals of varying ages and weights were extracted using a simple sample preparation procedure based on salt assisted liquid–liquid extraction. Data acquisition was performed using liquid chromatography and Q-Exactive™ Orbitrap mass spectrometry. Data processing and treatment were performed using two non-targeted workflows: (1) Compound Discoverer software and (2) XCMS package on the open-source R software combined with MetaboAnalyst. Three potential biomarkers were highlighted taking into account the chromatographic shapes, the feature location on the generated s-plots, the fold change, the adjusted p values, the coefficient of variation in the QC samples and the area under the ROC curves. Predicted formulas based on mass accuracy, structural composition and spectra similarity were proposed. A robust statistical model to predict the boldenone treatment was further developed based on the weighted abundances of the selected biomarkers. The requirements for screening methods were successfully fulfilled, together with a wider detection window in comparison with the monitoring of the deconjugated metabolite boldenone, although biomarker identification studies are still ongoing.

中文翻译：

---

使用超高效液相色谱与高分辨率轨道阱质谱联用检测牛中宝丹酮十一烯酸滥用的新策略：从非靶向到靶向

在这项工作中，采用了多变量策略，以突出新的潜在生物标志物，以检测用宝丹酮十一烯酸肌内治疗的公牛的外源性治疗。从处理（n  = 4）和对照（n = 8) 使用基于盐辅助液-液萃取的简单样品制备程序提取不同年龄和体重的杂交动物。使用液相色谱和 Q-Exactive™ Orbitrap 质谱进行数据采集。使用两个非靶向工作流程进行数据处理和处理：（1）Compound Discoverer 软件和（2）开源 R 软件上的 XCMS 软件包与 MetaboAnalyst 相结合。考虑到色谱形状、生成的 s 图上的特征位置、倍数变化、调整后的p值、QC 样品的变异系数和 ROC 曲线下的面积。提出了基于质量准确度、结构组成和光谱相似性的预测公式。基于所选生物标志物的加权丰度，进一步开发了预测宝丹酮治疗的稳健统计模型。尽管生物标志物鉴定研究仍在进行中，但与监测去结合代谢物宝丹酮相比，筛选方法的要求已成功满足，检测窗口更宽。