GeneSelectML

GeneSelectML selects differentially expressed genes from RNA-seq data using multiple machine learning algorithms to identify biologically relevant mRNA and miRNA candidates.

Key Features:

• Machine Learning Algorithms: Incorporates six distinct machine learning algorithms for gene selection.
• Pre-processing Capabilities: Includes filtering, normalization, transformation, and univariate analysis for RNA-seq data preparation.
• Inter-gene Relationship Analysis: Accounts for inter-gene relationships to improve predictive performance compared to traditional univariate analyses.
• Graphical Representations: Generates network plots, heatmaps, Venn diagrams, and box-and-whisker plots for data interpretation.
• Gene Ontology Analysis: Performs gene ontology analysis for both mRNA and miRNA differentially expressed genes.
• High-Performance Computation: Supports high-performance computation for processing large genomic datasets.

Scientific Applications:

• Alzheimer RNA-seq analysis: Applied to Alzheimer RNA-seq data and suggested eleven candidate genes including hsa-miR-148a-3p as a potential biomarker for Alzheimer's disease.
• Kidney Chromophobe validation: Validated using the Kidney Chromophobe dataset to demonstrate applicability across different genomic datasets.

Methodology:

The methodology involves simultaneous application of multiple machine learning algorithms to RNA-seq data, integrating classical pre-processing steps (filtering, normalization, transformation, univariate analysis) with advanced analytical techniques.