Hi-LASSO

Hi-LASSO applies a refined LASSO algorithm implemented for Python and Apache Spark to perform feature selection in high-dimensional biological datasets using parallel computing and parametric statistical tests to improve selection accuracy.

Key Features:

• High-Performance Parallel Computing: Implements parallel processing in Python and Apache Spark to reduce computation time for high-dimensional datasets.
• Superior Feature Selection: Demonstrates improved feature selection performance over other LASSO variants, aiding identification of relevant genes or biomarkers.
• Parametric Statistical Tests: Incorporates parametric statistical tests to validate selected features and provide a rigorous statistical framework for selection.
• Large-Scale Data Processing: Integrates with Apache Spark to support processing of large-scale bioinformatics datasets.

Scientific Applications:

• Genomics: Selection of relevant genes and genetic markers from high-dimensional genomic datasets.
• Proteomics: Identification of proteomic features and candidate protein biomarkers in high-dimensional proteomics data.
• Omics biomarker discovery: Extraction of meaningful patterns and biomarkers associated with diseases, genetic traits, or environmental interactions across omics studies.

Methodology:

Uses a refined LASSO (Least Absolute Shrinkage and Selection Operator) algorithm optimized for high-dimensional datasets combined with parallel processing and parametric statistical tests, provided as a Python package and an Apache Spark library.