TwoLocusGPR provides Gaussian-process regression utilities for two-locus tracking data. It bundles JAX-powered samplers, Gaussian-process marginal likelihoods, MSD-fitting routines, posterior geometry helpers, and example notebooks for downstream analysis.
pip install .Note that for GPU functionality, you need jax installed correctly, which is best set up before installing the package.
The workings of the module can be seen from the example notebooks:
- Examples/Anomaly_detection.ipynb: Detecting anomalies in time-series.
- Examples/fit_MSD.ipynb: Fitting MSDs.
- Examples/Identifying_Pol2_loadings.ipynb: Event-level polymerase inference.
- Examples/Posterior_predictions.ipynb: Predicting posteriors.