bio-machine-learning-biomarker-discovery

What problem does it solve?

High-dimensional omics data often contain many noisy features that hinder biomarker discovery. This skill guides users to identify informative features using Boruta all-relevant selection, mRMR minimum redundancy, and LASSO regularization to build robust biomarker panels.

Core Features & Use Cases

• Boruta all-relevant feature selection to capture all meaningful biomarkers.
• mRMR to minimize redundancy while maximizing relevance for compact biomarker panels.
• LASSO regularization to produce sparse, interpretable feature sets.
• Univariate pre-filtering to accelerate analysis on large feature spaces.
• Stability selection across bootstrap samples to identify robust biomarkers.
• Use Case: In a cancer omics dataset with thousands of genes, generate a biomarker panel for disease classification.

Quick Start

Run a full biomarker feature-selection workflow on your expression matrix X and labels y to produce a ranked biomarker list.