GeneSrF

GeneSrF implements a validated method for selecting very small sets of genes while preserving classification accuracy in the context of patient classification using microarray data.

Key Features:

• Random Forest Gene Selection: Implements a validated method for selecting minimal gene sets for patient classification from microarray data.
• Parallelized Computation: Computation is parallelized to take advantage of multicore CPUs and clusters of workstations.
• Bootstrap Error Estimation: Provides bootstrapped estimates of prediction error rate and assessments of solution stability.
• Functional Annotation Integration: Clickable tables link to GO terms, PubMed citations, and KEGG pathways for each selected gene.
• Pathway Analysis Export: Output can be sent to PaLS for examination of PubMed references, GO terms, KEGG and Reactome pathways characteristic of selected gene sets.

Scientific Applications:

• Patient Classification: Supports classification tasks using microarray gene expression data.
• Gene Selection: Identifies very small sets of genes that preserve classification accuracy for biomarker discovery.
• Exploratory Microarray Analysis: Enables applied biomedical researchers to perform rigorous exploratory work with microarray data while avoiding selection biases.

Methodology:

Gene selection and classification using random forest, with parallelized computation and bootstrap estimates of classification error rate.