3DMolMS

Version v1.3.2

3DMolMS (3D Molecular Network for Mass Spectra Prediction) is a deep-learning model that predicts the tandem mass (MS/MS) spectra of small molecules from their 3D conformations. Its learned molecular representation also transfers to related tasks such as retention time (RT) and collision cross section (CCS) prediction.

Use 3DMolMS to:

• Predict MS/MS spectra for small molecules

• Predict molecular properties — retention time (RT) and collision cross section (CCS)

• Generate reference MS/MS libraries for compound identification

• Pretrain models on 3D molecular datasets

There are two ways to run it: the molnetpack Python package for minimal-code inference and training, or the command-line source code for full control. New to the input files? Start with Supported formats.

PyPI package

• PyPI package: installation and inference
  • Installing from PyPI
  • Using molnetpack for MS/MS prediction
  • Plot predicted MS/MS
  • Using molnetpack for properties prediction
  • Molecular feature embedding
• PyPI package: training
  • train()
  • evaluate()
  • MS/MS model training
  • Retention time model training
  • CCS model training

Background

• Intro to MolConv
• Supported formats

Source code

• Source code setup
  • Installing from source code
  • Update package locally
  • Requirements
• Source code usage
  • Tandem mass spectra prediction
  • Train your own model for MS/MS prediction
  • Retention time model training
  • Collision cross section model training
  • Fine-tune on your own data
• Advanced features
  • Pretraining 3DMolMS on QM9
  • Generating a reference library for molecular identification

References

[1] Hong, Y., Li, S., Welch, C.J., Tichy, S., Ye, Y. and Tang, H., 2023. 3DMolMS: prediction of tandem mass spectra from 3D molecular conformations. Bioinformatics, 39(6), p.btad354.

[2] Hong, Y., Welch, C.J., Piras, P. and Tang, H., 2024. Enhanced structure-based prediction of chiral stationary phases for chromatographic enantioseparation from 3D molecular conformations. Analytical Chemistry, 96(6), pp.2351-2359.