CancerDetectionGPU

A GPU-accelerated tool for cell-free DNA based cancer diagnosis and tissue-of-origin prediction, based on the methods from the CancerLocator tool.

Prerequisites

• C++ compiler with C++17 support
• C++ compiler with support for OpenMP
• CMake (version 3.20 or higher)
• CUDA Toolkit (version 12.0 or higher)
• The GSL library (GNU Scientific Library)
• The BetaGPU library

Building the Project

The project uses CMake as its build system. To build the project, follow these steps:

cmake -S . -B build
cmake --build build --config Release

After building the project, the executable will be located in the build/bin directory.

Usage

You can run the program using the following command:

./CancerDetectionGPU <options>

Command-Line Options

Usage: CancerDetectionGPU [OPTIONS]

Options:
  -i, -input           Training data file [required] 
  -m, -test_methy      Test methylation data file [required] 
  -d, -test_depth      Test depth data file [required] 
  -t, -tissue_mapping  Tissue type mapping file [required] 
  -o, -output          Results file [required] 
  -r, -ratio_cut       Ratio cut 
  -c, -range_cut       Range cut 
  -n, -num_threads     Number of threads 
  -v, -verbose         Verbose 

Help options:
  -?, -help            print this help message 
      -about           print about message 

PROGRAM: CancerDetectionGPU
A tool for cell-free DNA based cancer diagnosis and tissue-of-origin prediction.

Example

An example command to run the program:

./bin/CancerDetectionGPU \ 
-i example/example_train \ 
-m example/example_test_methy \ 
-d example/example_test_depth \ 
-t example/example_type2class \ 
-o example/example_output \ 
-n 8 -v

Contact

[email protected]