Comparative Evaluation of DDA Strategies for Ginsenoside Profiling

The performance of three DDA approaches (M1 to M3) in acquiring the collision-induced dissociation (CID)-MS² data of PJM was compared. The general settings in Auto MS/MS (M1) were common for all three methods, and their only difference was the absence or inclusion of different PILs. In M1, no PIL was input, and intensity ranking-based selection of top 3 most intense precursors was achievable. M2 and M3 could be regarded as the PIL-including improved DDA strategies. For M2, the target masses (including 305 different m/z values, in total) corresponding to the known 579 ginsenosides, by considering the different adduct forms (e.g. both [M−H]⁻ and [M−H + HCOOH]⁻ for the neutral ginsenosides; [M−H]⁻ for the acidic saponins such as the OA-type and malonylated), were input. For M3, the target masses (103 different m/z values; Additional file 1: Table S2) resulting from the screening of the high-accuracy MS¹ data of PJM with MDF were input.
Development of the “Ginsenoside Sieve” was generally consistent with our previous report [44 (link)], which was based on the fixed variation range MDF and the in-house ginsenoside library. In detail, these 579 ginsenosides collected in the in-house database were in accordance with 185 different masses after removing the repeated values. The integer mass and decimal mass were distinguished by using the mod and trunc functions of Excel. The variation range, {Decimal mass − 10 mDa, Decimal mass + 10 mDa}, combined with the integer mass, could generate a sieve for ginsenosides. The established “Ginsenoside Sieve” was utilized to screen target m/z values from the MS¹ raw data of PJM processed by the MassHunter Workstation software.