omicsmad 0.1.1

A robust Scikit-Learn style Feature Selector using Median Absolute Deviation (MAD) for high-throughput genomics data.

omicsmad

A clean, minimal package for filtering genomics datasets (UCSC Xena bulk expression, CNV arrays) using median absolute deviation.

Installation

pip install omicsmad

Usage

import numpy as np
import pandas as pd
from omicsmad import OmicsMADSelector

# mocking data matrices
np.random.seed(42)
X_train = pd.DataFrame(np.random.randn(100, 5000), columns=[f"Gene_{i}" for i in range(5000)])

selector = OmicsMADSelector(top_n=1000)
selector.fit(X_train)

# saves to 'figures/mad_distribution.png'
selector.plot_distributions()

# saves processed matrix to 'datasets/processed/filtered_features.tsv'
selector.save_filtered_data(X_train)