基于离子化效率预测的液相色谱-高分辨质谱非靶向半定量分析研究进展OA

Non-targeted analytical methods based on liquid chromatography-high resolution mass spectrometry(LC-HRMS)have advanced rapidly in recent years,and are widely used in the fields such as food safety and environmental monitoring.However,quantification without reference standards remains a major challenge.To address this issue,establishing a quantitative structure-property relationship(QSPR)between ionization efficiency(IE)and molecular structure using multivariate linear regression or machine learning allows for prediction of IE for analytes lacking standards,thereby enabling non-targeted semi-quantitative estimation.This review summarized recent advances in IE prediction approaches for non-targeted semi-quantitative analysis using LC-HRMS,and discussed factors influencing electrospray ionization efficiency,the development of IE models and their applications in non-targeted semi-quantification.Strategies including data preparation,descriptor types and selection,multivariate linear regression and machine learning algorithms,and model calibration were described.Current limitations and challenges were discussed,and future perspectives on improving prediction accuracy and cross-platform applicability were presented.