Use analogy with Landau theory of phase transitions to look for critical transitions in data.
Associated publication:
Identifying a developmental transition in honey bees using gene expression data
Bryan C. Daniels, Ying Wang, Robert E. Page Jr, Gro V. Amdam
PLoS Computational Biology 19(9): e1010704; doi: https://doi.org/10.1371/journal.pcbi.1010704
The figures in the publication can be reproduced using the following notebooks:
- Figures 1, 4, 8: landau-simulation-analysis.ipynb
- Figures 2, 7: landau-bee-data-analysis.ipynb
- Figure 3, Table 1: LandauTransition-PaperDataAnalysis.nb
- Figure 5: critical-state-correlations.ipynb
- Figure 6A: hepatocellular-carcinoma-data.ipynb
- Figure 6B: LandauTransition-HCC.nb
- Python 3
- WolframScript (either through a Mathematica install or standalone)
- numpy
- scipy
- scikit-learn
- optional for running simulations: pandas, numba
To install the package and its python dependencies using pip
, clone the repository, descend into the landau/
folder, and run
pip install -e .
To automatically include the optional dependencies for running test simulations, instead run
pip install -e '.[simulation]'
WolframScript must also be installed on the system. This can be accomplished by installing Mathematica (and using the prepackaged Extras installer on Mac) or installed as a standalone package available for free download here: https://www.wolfram.com/wolframscript/