Application note: TDbasedUFE and TDbasedUFEadv: bioconductor packages to perform tensor decomposition based...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Tensor decomposition (TD)-based unsupervised feature extraction (FE) is effective for bioinformatics tasks like biomarker identification and gene analysis.
• Widespread adoption of TD-based FE has been limited by a lack of user-friendly tools for researchers without specialized expertise.

Purpose of the Study:

• To develop accessible tools for TD-based unsupervised FE in bioinformatics.
• To enable researchers unfamiliar with TD to perform advanced analyses such as differential gene expression and multiomics analysis.

Main Methods:

• Development of two R/Bioconductor packages: TDbasedUFE and TDbasedUFEadv.
• Implementation of TD-based unsupervised FE algorithms within user-friendly interfaces.
• Facilitation of differential gene expression analysis and multiomics data integration.

Main Results:

• The developed packages, TDbasedUFE and TDbasedUFEadv, successfully enable TD-based unsupervised FE for non-experts.
• TDbasedUFE demonstrated superior performance compared to state-of-the-art methods like DESeq2 and DIABLO in relevant analyses.
• The packages facilitate key bioinformatics tasks including identification of differentially expressed genes and multiomics analysis.

Conclusions:

• TDbasedUFE and TDbasedUFEadv lower the barrier to entry for utilizing powerful TD-based FE methods in bioinformatics research.
• These packages offer a robust and user-friendly alternative for feature extraction and analysis in genomics and multiomics studies.
• The tools are freely available, promoting broader application and advancement in the field.