Extracting Structural Information from Physicochemical Property Measurements Using Machine Learning─A New Approach for Structure Elucidation in Non-targeted Analysis.
Extracting Structural Information from Physicochemical Property Measurements Using Machine Learning─A New Approach for Structure Elucidation in Non-targeted Analysis.
复制标题
利用机器学习从理化性质测量中提取结构信息─ ─非靶向分析中结构解析的新方法。
DOI：
10.1021/acs.est.3c03003
复制
发表时间：
2023-10-10
影响因子：
11.4
通讯作者：
Woodruff, Tracey J.
中科院分区：
文献类型：
--
作者：
Abrahamsson, Dimitri;Brueck, Christopher L.;Prasse, Carsten;Lambropoulou, Dimitra A.;Koronaiou, Lelouda-Athanasia;Wang, Miaomiao;Park, June-Soo;Woodruff, Tracey J.
非靶向分析（NTA）在环境化学和环境健康领域做出了重要贡献。一个关键的瓶颈是缺乏针对环境中大多数化学品的分析标准。我们的研究旨在探索一种新的方法，将有机溶剂和水（KSW）之间的平衡分配比的测量，以预测分子结构。这些属性可以用作指纹，在机器学习算法的帮助下，可以将其转换为一系列官能团（RDKit片段），这些官能团可用于搜索化学数据库。我们进行了分配实验，使用的化学混合物含有185种化学品在10种不同的有机溶剂和水。使用液相色谱四极杆飞行时间质谱仪（LC-QTOF MS）和LC-Orbitrap MS来评估实验方法的可行性和预测正确官能团的算法的准确性。这两种方法在log KSW方面存在差异，QTOF方法的平均绝对误差（MAE）为0.22，Orbitrap方法为0.33。这些差异也导致RDKit片段预测的错误，QTOF方法的MAE为0.23，Orbitrap方法的MAE为0.31。我们的方法为NTA的结构解析提供了一个新的角度，并显示出协助化合物鉴定的希望。缺乏针对环境中大多数化学品的商业分析标准，是检测和测量环境和人体中有机污染物的一个关键瓶颈。本研究提出了一种新的方法，用于使用机器学习从物理化学性质测量中提取结构信息，用于通过非靶向分析检测到的化学品。
Non-targeted analysis (NTA) has made critical contributions in the fields of environmental chemistry and environmental health. One critical bottleneck is the lack of available analytical standards for most chemicals in the environment. Our study aims to explore a novel approach that integrates measurements of equilibrium partition ratios between organic solvents and water (KSW) to predictions of molecular structures. These properties can be used as a fingerprint, which with the help of a machine learning algorithm can be converted into a series of functional groups (RDKit fragments), which can be used to search chemical databases. We conducted partitioning experiments using a chemical mixture containing 185 chemicals in 10 different organic solvents and water. Both a liquid chromatography quadrupole time-of-flight mass spectrometer (LC-QTOF MS) and a LC-Orbitrap MS were used to assess the feasibility of the experimental method and the accuracy of the algorithm at predicting the correct functional groups. The two methods showed differences in log KSW with the QTOF method showing a mean absolute error (MAE) of 0.22 and the Orbitrap method 0.33. The differences also culminated into errors in the predictions of RDKit fragments with the MAE for the QTOF method being 0.23 and for the Orbitrap method being 0.31. Our approach presents a new angle in structure elucidation for NTA and showed promise in assisting with compound identification. The lack of commercially available analytical standards for most chemicals in the environment presents a critical bottleneck in detecting and measuring organic contaminants in the environment and in humans. This study presents a novel approach for extracting structural information from physicochemical property measurements using machine learning for chemicals detected through non-targeted analysis.
登录
查看更多内容
影响因子：
10.9
作者：
Plassmann, Merle M.;Fischer, Stellan;Benskin, Jonathan P.
通讯作者：
Benskin, Jonathan P.
影响因子：
4.3
作者：
Sobus, Jon R.;Grossman, Jarod N.;Ulrich, Elin M.
通讯作者：
Ulrich, Elin M.
影响因子：
3.6
作者：
Ganna, Andrea;Fall, Tove;Ingelsson, Erik
通讯作者：
Ingelsson, Erik
影响因子：
2
作者：
Kaufmann, Anton;Walker, Stephan
通讯作者：
Walker, Stephan
影响因子：
4.3
作者：
McEachran, Andrew D.;Sobus, Jon R.;Williams, Antony J.
通讯作者：
Williams, Antony J.