mixomics

mixomics performs multivariate analysis and integration of heterogeneous omics datasets (e.g., transcriptomics, proteomics, metabolomics) to identify correlated molecular signatures and discriminate biological conditions.

Key Features:

• Multivariate Analysis: Explores complex relationships within and between multiple omics datasets such as transcriptomics, proteomics, and metabolomics.
• Data Integration: Integrates various types of omics data measured on the same samples to enable cross-level biological analysis.
• Variable Selection: Implements simultaneous variable selection in integrated datasets to identify relevant molecular features.
• Dimension Reduction and Visualization: Provides dimension reduction techniques and visualization methods to summarize and inspect high-dimensional data structure.

Scientific Applications:

• Integrative Analysis: Uses regularized canonical correlation analysis and sparse partial least squares (PLS) regression to elucidate correlations between datasets.
• Discriminant Analysis: Extends PLS models for discriminant analysis to identify molecular signatures that explain or predict biological conditions across multiple omics datasets or independent studies.

Methodology:

Methods explicitly include regularized canonical correlation analysis, sparse partial least squares (PLS) regression, and Projection to Latent Structure (PLS) models extended for discriminant analysis.