⚠️ Experimental and under active development. Expect instability, breaking API changes, and incomplete features. No support or warranty. Use at your own risk.
Advanced NMR Peak Analysis and Integration Suite
PySide6/Qt6 suite for NMR peak detection, Voigt profile fitting, and integration — extended with binding-affinity (Kd / titration) fitting and protein dynamics (T₁/T₂, methyl T₂, spectral density, CPMG). Focus on automation: from automatic peak detection and overlap-aware multi-peak fitting through end-to-end batch processing of multi-spectrum datasets.
- Voigt fitting: Gaussian + Lorentzian lineshapes, ArPLS baseline correction, multi-peak deconvolution with simultaneous fitting of overlapping peaks.
- 2D simultaneous multi-peak fitting (PS2D): closely-spaced peaks auto-routed to 2D Levenberg-Marquardt deconvolution.
- Peak detection: network-based clustering with prominence analysis, no manual seeding required.
- Quality assessment: R², uncertainties (bootstrap / covariance), color-coded markers.
- Series workflows: relaxation (T₁, T₂, hetNOE), titration, and multi-spectrum (temperature/pH) processing with automated peak tracking.
- Kd / titration: CSP (quadratic 1:1 isotherm) and intensity-decay binding fits.
- Dynamics (DynamiXs): T₁/T₂, methyl T₂ (bi-exponential), spectral density, model-free, CPMG.
- Parallel processing: two-pass cluster-based, ~2.7× speedup.
- ML/statistics: dual-path fit-parameter prediction with statistical fallback (optional CNN classifier).
- Headless CLI:
python -m lunaNMRwithseries,dynamixs,kd,export,project,batchsubcommands — see docs/CLI.md.
File formats: Bruker TopSpin (.2ii, .2rr), Varian/Agilent (.fid, .ft), NMRPipe (.ft, .pipe), SPARKY (.ucsf). Export: CSV, JSON, PNG, PDF, SVG, EPS.
git clone https://github.com/mas-gu/lunaNMR.git
cd lunaNMR/lunaNMR_v1o0
pip install -r requirements.txt # dependencies
python3 lunaNMR/validation/verify_installation.py # verify
python3 launch_lunaNMR.py # GUI launcher (LunaNMR / DynamiXs selector)
python -m lunaNMR --help # headless CLIRequires numba (2D multi-peak fitting) and PySide6 (GUI). torch/torchvision optional (CNN classifier).
lunaNMR_v1o0/
├── launch_lunaNMR.py # GUI launcher (tkinter selector → Qt app)
├── lunaNMR/
│ ├── cli.py / __main__.py # headless CLI (python -m lunaNMR)
│ ├── core/ # peak picker, Voigt fitter, integrator, PS2D, parallel processor
│ ├── gui/ # PySide6/Qt6 interface
│ ├── processors/ # single/multi-spectrum workflows
│ ├── batch_processing/ # CLI batch automation
│ ├── ml/ # ML/statistics parameter prediction
│ ├── integrators/ # specialized integrators
│ ├── utils/ # config, project bundles, output routing
│ └── validation/ # installation verification
├── modules/dynamiXs_v2o0/ # DynamiXs (dynamics) + Kd / titration
└── docs/ # CLI.md, ARCHITECTURE.md, ALGORITHMS.md, guide
- Voigt function: convolution of Gaussian (instrumental) and Lorentzian (natural linewidth) components; multi-peak deconvolution with constraints.
- Peak detection: network-based clustering with prominence analysis.
- Optimization: Levenberg-Marquardt with parameter normalization.
- Uncertainty: bootstrap and covariance-based error estimation.
python3 -m venv dev_env && source dev_env/bin/activate # or dev_env\Scripts\activate on Windows
pip install -r requirements.txt
python -m pytest tests/Fork → feature branch → add tests → PR. PEP 8, NumPy-style docstrings.
@software{lunaNMR2025,
title = {LunaNMR: Advanced NMR Peak Analysis and Integration Suite},
author = {LunaNMR Contributors},
year = {2025},
version = {1.0},
url = {https://github.com/mas-gu/lunaNMR},
license = {MIT}
}MIT — see LICENSE.