The GEPIA website is freely available to all users. It is built by the HTML5 and JavaScript libraries, including jQuery (http://jquery.com), Bootstrap (http://getbootstrap.com/) for the client-side user interface. The server-side and interactive data processing are carried out by PHP scripts (version 7.0.13). The web site automatically adjusts the look and feel according to different browsers and devices, ranging from desktop computers to tablets and smart phones. There is no login requirement for accessing any features in GEPIA.
To solve the imbalance between the tumor and normal data which can cause inefficiency in various differential analyses, we download the TCGA and GTEx gene expression data that are re-computed from raw RNA-Seq data by the UCSC Xena project based on a uniform pipeline (Figure 1). We consult with medical experts to determine the most appropriate sample grouping for tumor-normal comparisons. The datasets are stored in a MySQL relational database (version 5.7.17).
The GEPIA web server features are divided into seven major tabs: General, Differential Genes, Expression DIY, Survival, Similar Genes, Correlation and PCA, which provides key interactive functions corresponding to differential expression analysis, customizable profiling plotting, patient survival analysis, similar gene detection, correlation analysis and dimensionality reduction analysis (Figure 2).
All plotting features in GEPIA are developed using R (version 3.3.2) and Perl (version 5.22.1) programs. The GEPIA outputs consist of plots and tables. Static visualizations are rendered as Portable Document Format (PDF), Scalable Vector Graphics (SVG) and Portable Network Graphics (PNG) images. The rotatable 3D plots are built by the plotly.js library (https://plot.ly/). Tables are generated by the DataTables (https://www.datatables.net/) JavaScript library, allowing for data querying and selection.