Novel Strategy for Mining and Identification of Acylcarnitines Using Data-Independent-Acquisition-Based Retention Time Prediction Modeling and Pseudo-Characteristic Fragmentation Ion Matching

It
is a challenging work to screen, identify, and quantify acylcarnitines
in complex biological samples. A method, based on the retention time
(RT) prediction and data-independent acquisition strategies, was proposed
for the large-scale identification of acylcarnitines using liquid
chromatography coupled with high-resolution mass spectrometry (LC-HRMS).
Relative cumulative eluotropic strength was introduced as a novel
descriptor in building a linear prediction model, which not only solves
the problem that acylcarnitines with long carbon chains cannot be
well predicted in traditional models but also proves its robustness
and transferability across instruments in two data sets that were
acquired in distinct chromatography conditions. The accessibility
of both predictive RT and MS2 spectra of suspect features
effectively reduced about 30% false-positive results, and consequently,
150 and 186 acylcarnitines were identified in the rat liver and human
plasma (NIST SRM 1950), respectively. This method provides a new approach
in large-scale analysis of acylcarnitine in lipidomic studies and
can also be extended to the analysis of other lipids.

在复杂生物样本中筛选、识别并定量酰基肉碱（acylcarnitines）是一项极具挑战性的工作。本研究提出一种基于保留时间（retention time, RT）预测与数据非依赖性采集策略的液相色谱-高分辨质谱联用（liquid chromatography coupled with high-resolution mass spectrometry, LC-HRMS）方法，用于大规模识别酰基肉碱。研究引入相对累积洗脱强度作为新型描述符构建线性预测模型，该模型不仅解决了传统模型无法对长碳链酰基肉碱实现精准预测的难题，还在两种不同色谱条件下获取的数据集间验证了其稳健性与仪器间可迁移性。通过有效利用可疑特征的预测保留时间与二级质谱（MS2）谱图，本研究将假阳性结果降低约30%，最终分别在大鼠肝脏样本与人类血浆（NIST SRM 1950）中鉴定出150种和186种酰基肉碱。该方法为脂质组学研究中酰基肉碱的大规模分析提供了全新途径，同时也可拓展至其他脂质类物质的分析。