This web service provies Patient-Based Query and Data Integration (PBQDI) for the geWorkbench.
- tumor type
- the name of sample file
- the content of the same file
- tumor type
- class assignments (map from sample names to subtype ID's)
- result package as a zipped package
- one HTML file
- one PDF file
- a number of image files. They are referenced from four different sections of the HTML file (the 'report'): one group for data quality, and the three groups for different categories of drugs, namely ontological, non-oncological, and investigational. The same image may show up in more than one categories of the drugs.
For this service to run, you need R installed, a number of required libraries, and the following R scripts and starting data files that are not included in the github repository due to copyright difference:
- classifySamples.r
- rununsupervised.r
- properties.r
- data-load-qc.r
- norm-viper.r
- oncoTarget-analysis.r
- tumorSubtypes.txt
- unsupervised.Rnw
- R version needs to be 3.3 or later
- 5 local packages need to be copied to the R library directory (specific location depending on your R installation). Their names are clinicalTrials, drugbank, n1database, n1platform, and tth. The actual code is not open-sourced.
- Biobase: used by n1platform. It is not avaiable in the public repository. To install, see https://bioconductor.org/packages/release/bioc/html/Biobase.html (No everything works forthrightly from this site. I ended up fiddling around more with BioInstaller to get it work on Linux.)
An example client program in Java is provided. The client does not store the ZIP file; instead, it unzips the contained files so they are ready to be used, e.g. the HTML file to be shown by a web browser or a browser component in other applications.