tspair

tspair identifies top scoring pairs of genes whose relative expression rankings classify samples into two predefined groups using rank-based comparison of gene expression data.


Key Features:

  • Simplicity and Interpretability: Classifiers use only two genes, enabling straightforward biological interpretation and facilitating development of assays such as PCR-based tests.
  • Robustness to Technical Variations: Rank-based classification uses relative gene rankings rather than absolute expression, reducing sensitivity to normalization and other technical variation.
  • Optimization of Sensitivity and Specificity: Identifies gene pairs that maximize the average of sensitivity and specificity across rank-based classifiers to assess candidate TSP classifiers.

Scientific Applications:

  • Disease Diagnosis: Identifies gene pairs that distinguish diseased and healthy states for potential diagnostic tests.
  • Biomarker Discovery: Pinpoints gene pairs with differential relative rankings across conditions to serve as biomarkers.
  • Personalized Medicine: Produces classifiers that stratify individuals based on gene expression profiles to inform treatment decisions.

Methodology:

Calculates pairs of genes whose relative rankings exhibit the maximum difference between two predefined groups and evaluates rank-based classifiers by optimizing the average of sensitivity and specificity to select top-scoring pairs.

Topics

Collections

Details

License:
GPL-2.0
Tool Type:
command-line tool, library
Operating Systems:
Linux, Windows, Mac
Programming Languages:
R
Added:
1/17/2017
Last Updated:
11/25/2024

Operations

Publications

Leek JT. The <tt>tspair</tt> package for finding top scoring pair classifiers in <tt>R</tt>. Bioinformatics. 2009;25(9):1203-1204. doi:10.1093/bioinformatics/btp126. PMID:19276151. PMCID:PMC2672632.

Documentation

Downloads