Related Experiment Video
Updated: Dec 7, 2025
09:38
Single-throughput Complementary High-resolution Analytical Techniques for Characterizing Complex Natural Organic Matter Mixtures
Published on: January 7, 2019
9.0K
MetFID: artificial neural network-based compound fingerprint prediction for metabolite annotation
Ziling Fan1, Amber Alley2, Kian Ghaffari2
1Department of Biochemistry and Molecular & Cellular Biology, Lombardi Comprehensive Cancer Center, Georgetown University Medical Center, Washington, DC, USA.
Metabolomics : Official Journal of the Metabolomic Society
|September 30, 2020
Summary
Metabolite annotation using mass spectrometry is improved by MetFID, a new method using artificial neural networks (ANNs) to predict molecular fingerprints. This enhances metabolite identification accuracy, especially for compounds lacking spectral data.
Area of Science:
- Metabolomics
- Computational Chemistry
- Bioinformatics
Background:
- Metabolite annotation in mass spectrometry-based metabolomics is challenging due to incomplete spectral libraries.
- Accurate metabolite identification is crucial for understanding biological pathways and disease mechanisms.
Purpose of the Study:
- To develop a novel computational method, MetFID, for ranking putative metabolite identities when experimental MS/MS spectra are unavailable.
- To improve the accuracy of metabolite annotation in untargeted metabolomic studies.
Main Methods:
- MetFID employs an artificial neural network (ANN) to predict molecular fingerprints from experimental MS/MS data.
- Candidate metabolites are retrieved from databases using precursor ion information (molecular formula or m/z).
- Annotation is achieved by matching predicted fingerprints with those of candidate metabolites.
Main Results:
- Training separate ANN models for different collision energy ranges improved performance.
- MetFID successfully prioritized correct metabolite IDs in the top rank for approximately 50% of test cases.
- MetFID demonstrated a >5% improvement in ranking accuracy compared to existing tools like ChemDistiller and MetFrag.
Conclusions:
- MetFID enhances metabolite annotation accuracy by leveraging ANNs for molecular fingerprint prediction.
- The method shows significant potential for improving metabolite identification in metabolomic research.
- MetFID offers a valuable tool for researchers dealing with incomplete spectral libraries.
