ISFrag

ISFrag identifies in-source fragments (ISFs) in LC-MS data to de novo recognize ISFs generated during electrospray ionization (ESI) and reduce false metabolite annotations in untargeted metabolomics.

Key Features:

• Pattern recognition: Detects ISF features by identifying three patterns: coelution with precursor ions, presence in MS² spectra of their precursor ions, and sharing similar MS² fragmentation patterns with their precursor ions.
• Acquisition-mode independence: Recognizes ISF features from LC-MS data regardless of acquisition mode, including full-scan, data-dependent acquisition (DDA), and data-independent acquisition (DIA).
• Library-agnostic detection: Operates without relying on common neutral loss information or pre-existing MS² spectral libraries.
• Accuracy and validation: Validation with metabolite standards reported 100% correct recognition for level 1 ISF features and over 80% accuracy for level 2 ISF features.
• Systematic parameter analysis: Enables investigation of how MS parameters—capillary voltage, end plate offset, ion energy, and collision energy—affect ISF generation and relative feature proportions.
• ISF pathway visualization: Generates an ISF pathway that visualizes relationships among multiple ISF features originating from the same precursor ion.

Scientific Applications:

• Untargeted metabolomics: Identifies and corrects potential false metabolite annotations caused by ISFs to improve annotation reliability.
• Omics-scale data interpretation: Enhances confidence in LC-MS–based metabolomic datasets by reducing ISF-driven annotation errors at scale.
• Pharmacology, toxicology, and systems biology: Supports more robust LC-MS–derived conclusions in these fields by improving metabolite identification fidelity.

Methodology:

Detects ISFs by three explicit criteria—coelution with precursor ions, presence in MS² spectra of the precursor, and shared MS² fragmentation patterns; processes LC-MS data across full-scan, DDA, and DIA modes without using common neutral loss information or pre-existing MS² spectral libraries; validated against metabolite standards (100% level 1, >80% level 2); and analyzes effects of capillary voltage, end plate offset, ion energy, and collision energy on ISF generation.