GeneSrF and varSelRF: a web-based tool and R package for gene selection and classification using random forest.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Microarray data analysis is crucial for patient classification and gene selection.
• Existing tools often lack user-friendliness, statistical rigor, or comprehensive functional gene information.
• There is a need for integrated solutions that avoid bias and offer multiple analysis options.

Purpose of the Study:

• To develop user-friendly yet statistically rigorous tools for gene selection from microarray data.
• To implement a validated method for identifying minimal gene sets while maintaining classification accuracy.
• To provide access to functional information for selected genes and make source code available.

Main Methods:

• Development of GeneSrF (web-based tool) and varSelRF (R package).
• Implementation of a validated gene selection method incorporating parallelized computation.
• Inclusion of bootstrapped estimates for prediction error rate and solution stability.

Main Results:

• GeneSrF and varSelRF successfully select very small gene sets with preserved classification accuracy.
• Tools offer parallelized computation for efficiency on multicore systems.
• Outputs include error rate estimates, stability assessments, and direct links to gene functional information (GO, PubMed, KEGG).

Conclusions:

• varSelRF and GeneSrF provide validated gene selection methods with bootstrap error estimates, ideal for exploratory microarray analysis.
• These tools are valuable for applied biomedical researchers.
• The combination of parallelization and web-based application offers methodological interest for bioinformaticians and biostatisticians.