Getting Started¶

CREDTOOLS is a command-line workflow for fine-mapping. You give it GWAS summary statistics and LD information. It gives you QC reports, posterior inclusion probabilities, and credible sets.

The docs are arranged around the question you probably have right now:

Which Starting Point Matches You?¶

I have raw GWAS filesI already have locus filesI only want one command

Start here if you have whole-genome summary statistics and genotype reference files such as PLINK .bed/.bim/.fam.

Create a population config file.
Run credtools munge.
Run credtools chunk.
Run credtools pipeline.

:octicons-arrow-right-24: Raw GWAS to Results

Start here if each locus already has:

{prefix}.sumstats.gz or {prefix}.sumstat
{prefix}.ld or {prefix}.ld.npz
{prefix}.ldmap or {prefix}.ldmap.gz

:octicons-arrow-right-24: Existing Loci List

Use credtools pipeline once your loci_list.txt is ready.

credtools pipeline loci_list.txt results --tool susie

The pipeline runs meta-analysis, QC, and fine-mapping.

The Mental Model¶

Think of the workflow as a series of small handoffs:

Step	What you have	What CREDTOOLS makes
`munge`	messy GWAS files	clean, standard summary statistics
`chunk`	clean genome-wide files + LD reference	locus-level files and a `loci_list.txt`
`meta`	one locus across cohorts or ancestries	combined data for fine-mapping
`qc`	summary statistics + LD	QC tables and outlier flags
`finemap`	one prepared locus set	PIPs and credible sets
`pipeline`	`loci_list.txt`	meta-analysis, QC, and fine-mapping in one run
`plot`	QC or result files	quick visual checks

A locus is the unit of work

CREDTOOLS fine-maps one genomic region at a time. If you start with whole-genome data, chunk creates those regions for you.

A Useful First Rule¶

Run a small example before a big batch. It catches path issues, missing LD files, and external tool problems before you spend hours on a full run.

credtools --help
credtools pipeline --help

Then continue to Installation or Quickstart.