Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances.

摘要: Biomarker detection is one of the more important biomedical questions for high-throughput 'omics' researchers, and almost all existing biomarker algorithms generate subset with optimized performance measurement a given dataset. However, recent study demonstrated existence multiple subsets similarly effective or even identical classification performances. This protocol presents simple straightforward methodology detecting binary performances, better than user-defined cutoff. The consists data preparation loading, baseline information summarization, parameter tuning, screening, result visualization interpretation, gene annotations, exportation at publication quality. proposed screening strategy intuitive demonstrates general rule developing algorithms. A user-friendly graphical user interface (GUI) was developed using programming language Python, allowing researchers to have direct access their results. source code manual kSolutionVis can be downloaded from http://www.healthinformaticslab.org/supp/resources.php.