A code for Pointwise Entropy Distributions for a Group Characteristic in High-dimensional and Sparse Microbiome Compositions
run main.py to try various simulation, you can change various conditions based on your preference.
You can test two types of simulation
- Same-ratio(k) Sensitivity Analysis
set parser.args.task
as 1 and by varying the level of same_ratio
and number of features(num_features
) you can check the p-value and its standard deviation after repeated permutation.
- Sparse-level Sensitivity Analysis
set parser.args.task
as 0 you are able to generate each group with various sparse ratio difference level, and you can check the p-value and its standard deviation.
from Example folder, you can see examples of synthetic data generation and meta analysis of them, , origianl version of permutation with phylum-level microbiome, and plots of ordered pointwise entropy. Instructions are detailed in Jupyter notebook.