easyMF

easyMF performs matrix factorization-based decomposition of large-scale RNA sequencing (RNA-Seq) transcriptome data to reduce dimensionality and identify metagenes for functional gene discovery and exploratory analysis.

Key Features:

• Matrix Preparation Module: Transforms raw reads from RNA-Seq experiments into processed gene expression data for downstream analysis.
• Matrix Factorization Module: Decomposes large-scale gene expression matrices into a smaller set of metagenes to reduce dimensionality and reveal underlying patterns.
• Metagene-based Deep Mining — Amplitude Matrix: Analyzes amplitude variations across metagenes to uncover functional insights.
• Metagene-based Deep Mining — Pattern Matrix: Utilizes pattern matrices to explore gene expression patterns and identify functionally relevant genes.

Scientific Applications:

• Dimensionality reduction and feature extraction: Applies to high-dimensional gene expression matrices generated by RNA-Seq to obtain compact metagene representations.
• Functional gene discovery: Enables identification of genes associated with metagenes and putative biological functions via amplitude and pattern analyses.
• Exploratory transcriptome analysis: Supports exploratory analysis of large-scale RNA-Seq datasets, demonstrated on 940 RNA sequencing datasets from maize (Zea mays L.).

Methodology:

Computational steps explicitly include transforming raw reads into processed gene expression matrices, applying matrix factorization to reduce thousands of genes into metagenes, and performing metagene-based deep mining using amplitude and pattern matrices.