Skip to content

credtools

pypi python Build Status codecov License: MIT

Multi-ancestry fine-mapping pipeline.

Features

  • Whole-genome preprocessing: Start from raw GWAS summary statistics and genotype data
  • Standardize and munge summary statistics from various formats
  • Prepare LD matrices and fine-mapping inputs automatically
  • Multi-ancestry fine-mapping: Support for multiple fine-mapping tools (SuSiE, FINEMAP, etc.)
  • Meta-analysis capabilities: Combine results across populations and cohorts
  • Quality control: Built-in QC metrics and visualizations
  • Command-line interface: Easy-to-use CLI for all operations

Installation

Basic Installation

pip install credtools

Install with uv

uv pip install credtools

Quick Start

Command Line Usage

# Complete workflow: from whole-genome data to fine-mapping results
# Step 1: Standardize summary statistics
credtools munge population_config.txt output/munged/

# Step 2: Identify loci, chunk data, and extract LD matrices
credtools chunk output/munged/sumstat_info_updated.txt output/chunks/

# Step 3: Run fine-mapping pipeline
credtools pipeline output/chunks/loci_list.txt output/results/

Preprocessing Workflow

credtools now supports starting from whole-genome summary statistics and genotype data, eliminating the need for manual preprocessing:

Step 1: Munge Summary Statistics (credtools munge)

  • Purpose: Standardize and clean GWAS summary statistics from various formats
  • Features:
  • Automatic header detection and mapping
  • Data validation and quality control
  • Support for multiple file formats
  • Input: Raw GWAS files with various column headers
  • Output: Standardized .munged.txt.gz files

Step 2: Chunk Loci (credtools chunk)

  • Purpose: Identify independent loci, create regional chunks, and extract LD matrices
  • Features:
  • Distance-based independent SNP identification
  • Cross-ancestry loci coordination
  • Configurable significance thresholds
  • Automatic LD matrix extraction when ld_ref is provided in population config
  • Input: Munged summary statistics files (or population config with ld_ref)
  • Output: Locus-specific chunked files, LD matrices, and credtools-ready input files

Multi-Ancestry Support

  • Consistent loci definition: Union approach across ancestries
  • Flexible input formats: Support for various GWAS summary statistics formats
  • Coordinated processing: Ensure compatibility across populations

Documentation

For detailed documentation, see https://Jianhua-Wang.github.io/credtools