Quantum Chemistry Calculations for Metabolomics

Ricardo M. Borges, Sean Colby, Susanta Das, Arthur S. Edison, Oliver Fiehn, Tobias Kind, Jesi Lee, Amy T. Merrill, Kenneth M. Merz, Thomas Metz, Jamie Nuñez, Dean J. Tantillo, Lee‐Ping Wang, Shunyang Wang, Ryan RenslowChemical 2021 · DOI

Abstract

A primary goal of metabolomics studies is to fully characterize the small-molecule composition of complex biological and environmental samples. However, despite advances in analytical technologies over the past two decades, the majority of small molecules in complex samples are not readily identifiable due to the immense structural and chemical diversity present within the metabolome. Current gold-standard identification methods rely on reference libraries built using authentic chemical materials ("standards"), which are not available for most molecules. Computational quantum chemistry methods, which can be used to calculate chemical properties that are then measured by analytical platforms, offer an alternative route for building reference libraries, i.e., in silico libraries for "standards-free" identification. In this review, we cover the major roadblocks currently facing metabolomics and discuss applications where quantum chemistry calculations offer a solution. Several successful examples for nuclear magnetic resonance spectroscopy, ion mobility spectrometry, infrared spectroscopy, and mass spectrometry methods are reviewed. Finally, we consider current best practices, sources of error, and provide an outlook for quantum chemistry calculations in metabolomics studies. We expect this review will inspire researchers in the field of small-molecule identification to accelerate adoption of in silico methods for generation of reference libraries and to add quantum chemistry calculations as another tool at their disposal to characterize complex samples.

Introduction: The "Dark Matter" of the Metabolome

Metabolomics faces a unique challenge compared to genomics or proteomics. While the latter work with a finite set of building blocks (nucleotides or amino acids), the metabolome encompasses a staggering diversity of small molecules, with estimates of possible structures exceeding $10^{60}$. Currently, the "gold standard" for identifying these molecules relies on comparing experimental signatures against libraries of physical "standards." However, physical standards exist for less than 10% of known metabolites, leaving a vast amount of "dark matter" in biological samples that cannot be identified.

The QM Solution: Standards-Free Identification

Quantum Chemistry (QM) offers a paradigm shift: "standards-free" identification. By using first-principles physics, specifically Density Functional Theory (DFT), researchers can predict the analytical signatures of any hypothetical molecule. This allows for the creation of in silico reference libraries—virtual standards that can be generated for molecules that have never been synthesized or isolated.

Quantum chemistry (QM) calculations enable the prediction of various analytical properties, including MS fragmentation, NMR, IR, and Collision Cross Section (CCS), facilitating standards-free metabolite identification.

Predicting Analytical Signatures

Computational methods are being refined across several key analytical modalities to bridge the gap between structure and spectrum:

  • NMR Spectroscopy: QM methods can predict chemical shifts and coupling constants with high precision. For instance, automated protocols for the folate molecule ($C_{19}H_{19}N_7O_6$) achieved a Mean Absolute Error (MAE) of just 0.26 ppm for $^1$H and 1.63 ppm for $^{13}$C, sufficient for high-confidence structural assignment.
  • Ion Mobility Spectrometry (IMS): By simulating the 3D geometry of ions, QM can calculate the Collision Cross Section (CCS), representing the molecule's "size" in a buffer gas. Hybrid ML/QM approaches have reduced calculation times from hours to milliseconds with minimal error.
  • Infrared (IR) Spectroscopy: Predicting vibrational frequencies allows for structural confirmation, though it often requires anharmonic corrections for high accuracy.
  • Mass Spectrometry (MS): This remains a significant challenge. Methods like QCEIMS (Quantum Chemistry Electron Ionization Mass Spectrometry) use molecular dynamics to simulate fragmentation patterns, providing a first-principles approach to interpreting complex MS/MS data.

The Synergy of Quantum Chemistry and Machine Learning

While QM is physically accurate, it is computationally expensive. The future of the field lies in hybrid QM/ML approaches. Machine learning models, trained on QM-generated data, can accelerate property prediction by orders of magnitude. For example, ML-based potentials can quickly filter through thousands of molecular conformations before performing expensive DFT calculations on the most likely candidates.

Challenges and Future Outlook

Significant hurdles remain, including the high computational cost for large molecules (like lipids), the need for better solvent and pH modeling, and the difficulty of predicting tautomeric mixtures. Real-world metabolites exist in complex biological fluids, not in a vacuum. Furthermore, while we are proficient at predicting a spectrum from a structure, the "inverse problem"—determining a structure directly from a spectrum—remains a major frontier for the next decade.

Conclusion

By integrating quantum chemistry into the metabolomics workflow, researchers can begin to illuminate the "dark matter" of the chemical world. Moving beyond the limitations of physical standards toward a truly comprehensive characterization of biological systems will require the continued evolution of both computational power and theoretical models.

Summary updated

Abstract

A primary goal of metabolomics studies is to fully characterize the small-molecule composition of complex biological and environmental samples. However, despite advances in analytical technologies over the past two decades, the majority of small molecules in complex samples are not readily identifiable due to the immense structural and chemical diversity present within the metabolome. Current gold-standard identification methods rely on reference libraries built using authentic chemical materials ("standards"), which are not available for most molecules. Computational quantum chemistry methods, which can be used to calculate chemical properties that are then measured by analytical platforms, offer an alternative route for building reference libraries, <i>i.e.</i>, <i>in silico</i> libraries for "standards-free" identification. In this review, we cover the major roadblocks currently facing metabolomics and discuss applications where quantum chemistry calculations offer a solution. Several successful examples for nuclear magnetic resonance spectroscopy, ion mobility spectrometry, infrared spectroscopy, and mass spectrometry methods are reviewed. Finally, we consider current best practices, sources of error, and provide an outlook for quantum chemistry calculations in metabolomics studies. We expect this review will inspire researchers in the field of small-molecule identification to accelerate adoption of <i>in silico</i> methods for generation of reference libraries and to add quantum chemistry calculations as another tool at their disposal to characterize complex samples.
BibTeX
@inproceedings{borges2021quantum,
  author = {Ricardo M. Borges and Sean Colby and Susanta Das and Arthur S. Edison and Oliver Fiehn and Tobias Kind and Jesi Lee and Amy T. Merrill and Kenneth M. Merz and Thomas Metz and Jamie Nuñez and Dean J. Tantillo and Lee‐Ping Wang and Shunyang Wang and Ryan Renslow},
  title = {{Quantum Chemistry Calculations for Metabolomics}},
  booktitle = {Chemical},
  year = {2021},
  doi = {10.1021/acs.chemrev.0c00901},
}
Figure