CFM-ID

CFM-ID predicts tandem mass spectrometry (MS/MS) spectra, annotates spectral peaks, and ranks candidate metabolites for compound identification using competitive fragmentation modeling.

Key Features:

• Spectral Peak Annotation: Assigns MS/MS spectral peaks to fragments derived from a specified chemical structure.
• MS/MS Spectrum Prediction: Predicts tandem mass spectra for given chemical structures based on fragmentation modeling.
• Metabolite Candidate Ranking: Generates ranked candidate molecular structures for a query spectrum to support metabolite identification.
• Machine Learning-Based Fragmentation Modeling: Applies trained probabilistic models to simulate chemical fragmentation patterns in mass spectrometry.

Scientific Applications:

• Metabolite Identification: Identifies unknown metabolites from tandem mass spectrometry data.
• Metabolomics Data Analysis: Interprets MS/MS spectra to characterize small molecules in metabolomic studies.

Methodology:

CFM-ID uses Competitive Fragmentation Modeling, a probabilistic generative machine learning model trained on mass spectrometry datasets, to simulate fragmentation processes and predict spectra or rank candidate molecular structures.