OSCA

OSCA (OmicS-data-based Complex trait Analysis) is a software tool for the analysis of complex traits using multi-omics data and genetic analysis of molecular phenotypes.

Requirement

• zlib >= 2.1

• gsl >= 2.6

• Eigen

• MKL(Math Kernel Library)

  MKL is distributed within intel oneAPI, you can get this by install intel oneAPI or install oneMKL conponent.

• Rmath library

Instatll requirements

• Eigen

  Eigen just contain head files, and do not need compile to install it. Here, we use version 3.3.7 when we compile osca.

• MKL

  You can read document and download MKL at this webpage: https://www.intel.com/content/www/us/en/developer/tools/oneapi/toolkits.html#base-kit.

• Rmath library

  Note: functions of Rmath are also contained by libR.so/libR.a, you can use them instead of Rmath to make life easier.

  I using version 3.6 as example to install this library.

  • download R

    wget https://cloud.r-project.org/src/base/R-3/R-3.6.3.tar.gz
tar -zxf R-3.6.3.tar.gz
cd R-3.6.3

  • configure

    ./configure --prefix="PATH where you what put this lib"

  • compile and install

    cd src/nmath/standalone
make
make install

  Then the head files and library files would show up under prefix path.

Compile

• Method 1, using Makefile

  Samplely, if all requirements is installed under /usr, just run following command, then the osca would be compiled.

  make  
  

  This would search head file under /usr/include and search library under /usr/lib64.

  If you don't install requirements in other directory, you may need edit Makefile or specific following variables when run make command.

  make EIGEN_PATH="eigen head file path" MKL_INCLUDE="mkl head file path" \
  MKL_LIB="mkl library file path" RMath_INCLUDE="path where install Rmath head files" RMath_LIB="path where install Rmath library"
  

  By default, only the dynamic version would compiled, if you want compile static version, run:

  make VIRABLE="VALUE" osca_static
  

  Using DEBUG=ON to switch on debug mode.

  After compilation, a file named osca would appared under same directory of Makefile, and if you compile static version, the executable file, named osca_static, would show up too.

• Method 2, using cmake

      git clone https://github.com/benjaminfang/osca.git
      cd osca
      mkdir build
      cd build
      cmake ..
      make
  

USAGE

See https://yanglab.westlake.edu.cn/software/osca/#Overview for its usage and data resources.

MCP Server — Use OSCA via AI Conversation

The pre-built binary directory (osca-1.22-linux-x86_64/) ships with an MCP server (mcp_server.py) that exposes all OSCA functionality to AI assistants such as Claude Desktop, OpenAI Codex, and opencode.

Quick Start

pip install mcp

Add the server to your AI tool's config (replace <OSCA_DIR> with the real path to the osca-1.22-linux-x86_64 directory):

{
  "mcpServers": {
    "osca": {
      "command": "python3",
      "args": ["<OSCA_DIR>/osca-mcp/mcp_server.py"]
    }
  }
}

Restart your AI tool, then ask questions in natural language:

• "Create an ORM from my methylation data"
• "Run a mixed linear model association analysis with covariates"
• "What flags does OSCA support for eQTL analysis?"

Available MCP Tools

Tool	Description
osca_info()	Check OSCA binary status and version
osca_help(topic)	Get OSCA documentation (overview, commands, flags, examples, formats)
run_osca(args, workdir)	Execute any OSCA command; returns stdout, stderr, output files, and log
list_files(path, pattern)	List files in a directory
read_file(path, max_lines)	Read file contents with pagination

See osca-1.22-linux-x86_64/Readme.txt for full configuration details.