GraphLD¶
GraphLD provides LDGM-backed tools for working with GWAS summary statistics.
You can use it from the command line for graphREML, BLUP, clumping, simulation, and surrogate-marker workflows, or from Python for lower-level LDGM and likelihood operations.
Start Here¶
Choose the path that matches what you want to do:
- If you have new annotations that you wish to test for enrichment against a prespecified set of traits, go to Enrichment Score Test.
- If you have new annotations or new GWAS summary statistics and wish to estimate heritability enrichments, go to CLI graphREML.
- If you need to install GraphLD or download reference data, start with Installation.
- If you want a broader overview of the terminal workflows, start with Command Line Interface.
- If you want to use GraphLD from Python, start with the Python Guide.
- If you need exact signatures for modules, classes, or functions, use the autogenerated API Reference.
- If you are unsure what file layout GraphLD expects, check File Formats.
Features¶
- graphREML for heritability partitioning and enrichment from GWAS summary statistics.
- Enrichment score tests for very fast heritability enrichment significance testing from precomputed derivative statistics at either a gene or variant level.
- Simulation for generating GWAS summary statistics with realistic LD patterns.
- BLUP weights for performing polygenic prediction.
- LD clumping for selecting lead associations or performing polygenic prediction.
- LDGM matrix operations for precision-matrix and Schur-complement operations.
- Surrogate markers for accelerating graphREML when you have several sets of summary statistics with similar variant missingness.
Documentation Map¶
- Installation: environment setup, dependencies, and data downloads
- Command Line Interface: shared CLI concepts plus per-subcommand pages
- Enrichment Score Test:
estestworkflows and HDF5 score-stat usage - Python Guide: workflow-oriented examples for using GraphLD from Python
- File Formats: expected summary-statistics, annotation, and LDGM inputs
- Reference overview: autogenerated module-level API documentation