Improving compound identification results by automatically recognizing in-source fragment ions in HRMS with AntDAS: A study on accurate pesticide screening in complex food samples.

Ultrahigh-performance liquid chromatography combined with high-resolution mass spectrometry has been extensively used for compound identification in an untargeted manner. However, current methods still cannot accurately screen and identify compounds in the presence of in-source fragment ions. Here, we proposed an in-source fragment ion-based strategy to precisely recognize compounds. First, raw data files were analyzed with our newly developed AntDAS-DDA to perform feature detection and annotation. Second, in-source fragment ions from a compound were recognized based on tandem mass spectrometry (MS/MS) spectra and the similarity of chromatograms. Last, the MS/MS spectra of each fragment ion were compared with those in an in-house library to benefit compound identification. The performance of the developed strategy was comprehensively investigated using the aforementioned in-house library containing 478 pesticides. Simulated and real samples suggested that compounds were accurately identified. The match factor obtained from the classic method was improved by at least 4.3 times by the developed strategy. The developed strategy may provide a new solution for compound identification in complex food samples.