KODAMA

KODAMA performs unsupervised feature extraction to identify structure in noisy, high-dimensional biological datasets.

Key Features:

• Unsupervised feature extraction: Implements methods to identify patterns without labeled data for exploratory analysis.
• Novel learning algorithm: Uses a specialized learning algorithm developed for robust feature extraction from complex datasets.
• Handling noisy data: Robust to substantial noise typical of biological and bioinformatics datasets, preserving meaningful signal.
• High-dimensional data analysis: Optimized for datasets with large numbers of variables such as genomics and proteomics profiles.
• Interpretability functions: Includes supplementary functions to aid interpretation of extracted features and high-dimensional structures.
• R package implementation: Provided as an R package for integration into computational workflows.

Scientific Applications:

• Gene expression analysis: Extracts features and patterns from bulk gene expression datasets.
• Single-cell RNA sequencing: Supports exploration and interpretation of single-cell RNA-seq data.
• Integrative multi-omics studies: Facilitates extraction of shared structure across multi-omics datasets.
• Exploratory analysis and hypothesis generation: Reveals hidden structures in biological data to support hypothesis generation.

Methodology:

KODAMA applies a novel unsupervised learning algorithm for feature extraction and provides additional functions for interpreting high-dimensional results.