Figure 1.

HFDiscrim: identification of candidate peak group using hybrid features. (A) Following the analysis of DIA mass spectrometry data by OpenSWATH and PyProphet, candidate peak group and their single-run subscores (SR subscores, SR-Subs) and single-run discriminant score (SRDS) are obtained. To mitigate intersample variability, the retention times across different runs are aligned. Subsequently, cross-run subscores are calculated based on the chromatographic peak signals from multiple samples. Finally, the identification results are discriminated based on the hybrid features. (B) The chromatographic peak with the highest SRDS in each sample is selected as the reference peak group. All candidate peak groups are then compared to the reference peak group to derive cross-run subscores (CR-Subs). (C) Hybrid features are composed of single-run features and cross-run features. For instance, two types of hybrid subscores were created for peak 11. (D) These hybrid subscores are then processed through a semisupervised model to yield a hybrid discriminant score (HDS). For peak 11, only the hybrid features corresponding to the highest HDS were retained.