Skip to content

synth-nmr

NMR spectroscopy calculations for protein structures

A lightweight, standalone Python package for calculating NMR observables from protein structures.

synth-nmr provides a focused toolkit that works with any protein structure source to predict experimental observables like NOEs, Relaxation rates, Chemical Shifts, and Residual Dipolar Couplings.

Features

  • NOE Calculations: Synthetic Nuclear Overhauser Effect distance restraints.
  • Relaxation Rates: \(R_{1}\), \(R_{2}\), and heteronuclear NOE predictions based on Model-Free formalism.
  • Chemical Shifts: High-accuracy Neural Network predictions with SPARTA+ empirical fallback.
  • J-Couplings: Karplus equation applications for scalar couplings.
  • RDC Calculations: Prediction of Residual Dipolar Couplings from alignment tensors.
  • NEF I/O: Native support for the NMR Exchange Format.

Why synth-nmr?

Modern structural biology increasingly relies on hybrid methods. Whether you're generating structures using AlphaFold, running Molecular Dynamics simulations, or building de novo generative models, synth-nmr enables you to quickly validate those theoretical structures against standard NMR experimental observables.

If you are an AI/ML researcher, bridging the gap between 3D Cartesian coordinates and the measurements that spectroscopists actually record is crucial for training multimodal models. synth-nmr acts as that rigorous biophysical bridge.

Next Steps

  • Installation & CLI Usage: Learn how to install the package and use the command-line interface.
  • Scientific Background: Dive deep into the physics and theory of NMR, including the seminal contributions of pioneers like Wüthrich and Bax.
  • API Reference: Browse the Python API documentation for integrating synth-nmr into your own pipelines.