SMR

The SMR software tool was originally developed to implement the SMR & HEIDI methods to test for pleiotropic association between the expression level of a gene and a complex trait of interest using summary-level data from GWAS and expression quantitative trait loci (eQTL) studies (Zhu et al. 2016 Nature Genetics). The SMR & HEIDI methodology can be interpreted as an analysis to test if the effect size of a SNP on the phenotype is mediated by gene expression. This tool can therefore be used to prioritize genes underlying GWAS hits for follow-up functional studies. The methods are applicable to all kinds of molecular QTL (xQTL) data, including DNA methylation QTL (mQTL) and protein abundance QTL (pQTL).

Requirement

• Eigen

• libz

Installation

Install requirments

Dependencies (Eigen, zlib, spdlog, fmt, mimalloc) are fetched automatically by CMake — no manual installation needed. A C++17 compiler, CMake ≥ 3.16, and OpenMP are required.

Build

using make

Simply, in smr directory (delegates to scripts/local/build.sh):

make

using cmake

In SMR directory,

mkdir build
cd build
cmake ..
make

The default build type is Release. To build with debug information,

cmake -DCMAKE_BUILD_TYPE=Debug ..

Useful CMake options:

• -DBUILD_WITH_MKL=ON/OFF (default ON): accelerate Eigen with Intel oneMKL; falls back to Eigen's native backend if MKL is not found. Use -DMKL_DIR=... or -DMKL_ROOT=... to point at your oneMKL installation.
• -DBUILD_WITH_MIMALLOC=ON/OFF (default ON): link mimalloc for faster memory allocation.
• -DSMR_VERBOSE_CONFIGURE=ON: print extended CMake configuration diagnostics.

Helper build scripts: scripts/local/build.sh [-f|--fresh] [-g|--generate] (Linux/macOS, uses oneMKL) and scripts/hpc/build.sh (HPC clusters).

Usage

visit https://yanglab.westlake.edu.cn/software/smr/ for software's document.

Using SMR with AI Tools

The SMR release package includes a built-in MCP (Model Context Protocol) server, enabling you to run SMR analyses through AI tools such as Claude, Codex, and OpenCode via natural language conversation.

Quick Start

1. Download and extract the SMR release package.
2. Configure the MCP server for your AI tool (see CLAUDE_CODE_USAGE.md in the release package for details).
3. Start chatting with the AI — for example: "Run an SMR analysis with bfile data/1kg_eur, gwas-summary data/bmi.ma, beqtl-summary data/eqtl.besd, output to results/smr_out".

Available MCP Tools

Tool	Function
run_smr_analysis	Run the main SMR test (Wald ratio + HEIDI)
smr_multi	Run set-based (multi-SNP) SMR analysis
make_besd	Create a sparse BESD file from a text eQTL summary
make_besd_dense	Create a dense BESD file from a text eQTL summary
query_besd	Query a BESD file for significant SNP-probe associations
show_sample_size	Display the sample size stored in a BESD file
recode_besd	Convert a BESD file to COJO/SMR text format
plot_smr	Generate SMR locus plots
make_bld	Create a BLD (binary LD) file from a PLINK bfile
update_freq	Update allele frequencies in a BESD file
meta_analysis	Run meta-analysis of multiple eQTL studies (MeCS)
combine_besd	Combine multiple BESD files into one
count_cis	Count cis-eQTL in a BESD file
count_trans	Count trans-eQTL in a BESD file
update_epi_esi	Update EPI/ESI annotation files in a BESD file
get_version	Get SMR version information
run_raw_command	Pass arbitrary SMR command-line arguments (escape hatch)

For full configuration instructions for Claude Desktop, Claude Code, OpenCode, and Codex, see CLAUDE_CODE_USAGE.md in the release package.