CancerVis: An interactive exploratory tool for cancer biomarker analysis

摘要

A biomarker is a measurable substance that can be used as an indicator of a particular disease. In the research field of cancer biomarker analysis, data visualization has generally been utilized as a technique for presenting results, with very little application of integrating data visualization as a tool for biological data exploration or pattern discovery. To address this issue, we have developed a visualization platform, CancerVis, for visual and interactive analysis of cancer data and pattern discovery. CancerVis provides multiple interactive views from different prospective of a dataset. The views are synchronized so that users can easily link the same data entry in multiple views. CancerVis's visualization technologies also supports common data mining techniques for pattern discovery in cancer biomarker research, such as visualization of optimal cutoff point and its variability over time. The usability of CancerVis is verified by a clinical dataset.