GSgalgoR

GSgalgoR implements a multi-objective optimization algorithm for disease sub-type discovery based on a non-dominated sorting genetic algorithm.

Key Features:

• Clustering for omics data: Groups heterogeneous 'omics' data using clustering algorithms.
• Genetic algorithm feature selection: Uses genetic algorithm searching properties for feature selection.
• Multi-objective optimization: Searches for the optimal number of clusters by maximizing survival difference between sub-types while maintaining high cluster consistency.

Scientific Applications:

• Cancer prognostic gene expression signature identification: Identifies and studies prognostic gene expression signatures in cancer through disease sub-type discovery.
• Survival-based patient stratification: Discovers patient sub-types with maximized survival differences for clinical outcome prediction.

Methodology:

The Galgo framework combines clustering algorithms for grouping heterogeneous omics data with genetic algorithms for feature selection, using non-dominated sorting to optimize cluster number determination.