Comprehensive Profiling of Microcystins in Non-Targeted LC-HR-MS Datasets using Mass Spectrometry Query Language (MassQL)

Microcystins (MC) are a critical class of toxic cyanobacterial secondary metabolites, with specific variants such as MC-LR strictly regulated in drinking water. Despite their extensive structural diversity, comprising over 300 known congeners, comprehensive environmental monitoring is severely hindered by the limited availability of commercial reference standards and public tandem mass spectrometry (MS/MS) spectra. To address this analytical bottleneck, we developed a non-targeted mass spectrometry-based computational workflow leveraging the Mass Spectrometry Query Language (MassQL) to enable comprehensive MC profiling. By systematically analyzing the chemical space and fragmentation patterns of all MC congeners documented in the CyanoMetDB database, we classified the congeners into nine structurally distinct groups. Subsequently, we designed and optimized a suite of MassQL queries targeting the conserved MC backbone, group-specific motifs, and individual constituent amino acids. The performance of these custom queries was rigorously evaluated following Feature-Based Molecular Networking (FBMN). Application across three independent cyanobacterial bloom datasets comprising both in-house and public data acquired via qTOF and Orbitrap high-resolution mass spectrometry platforms demonstrated highly satisfactory analytical performance based on recall, precision, F1-scores, and Matthews Correlation Coefficients (MCC). We further demonstrated that MassQL functions as a powerful MC candidate prioritization tool orthogonal to molecular networking. By interrogating MS/MS data for specific structural motifs, this query-based approach enabled the discovery and annotation of novel MC congeners lacking public library spectra, significantly enhancing non-targeted screening for water quality and risk assessment.