Statistical genetics · Software
GCTA
Tool for estimating heritability, genetic correlation, and running association tests from genome-wide data.
What GCTA is
GCTA (Genome-wide Complex Trait Analysis) is a statistical genetics package developed by Jian Yang and collaborators. Originally designed to estimate the proportion of phenotypic variance explained by all genome-wide single nucleotide polymorphisms (SNPs) for complex traits, it addresses the problem of quantifying 'missing heritability' in genome-wide association studies (GWAS).
Over time, the package has expanded to support a broad set of analyses for post-GWAS and biobank-scale data. Researchers use it to calculate genetic relationship matrices (GRMs), partition heritability across functional genomic annotations or chromosomes, compute genetic correlations between traits, and run computationally efficient association analyses across large cohorts.
How it works
GCTA constructs genetic relationship matrices from SNP datasets and applies methods such as genomic restricted maximum likelihood (GREML) and Haseman-Elston regression to estimate variance components and heritability. These models can incorporate fixed covariates, partition variance by chromosome or functional annotation, and evaluate dominance variance (GREMLd) or bivariate genetic correlations.
For association testing, GCTA implements mixed linear model association (MLMA and MLMA-LOCO) and scalable alternatives. Modules such as fastGWA and fastGWA-GLMM employ sparse genetic relationship matrices to conduct linear and generalized linear mixed-model association analyses on biobank-scale data. Additional modules, including COJO, mtCOJO, fastBAT, and GSMR, analyse summary statistics to perform multi-SNP conditional testing, gene-set tests, and generalised summary-data-based Mendelian randomisation.
Key facts
- Category
- Statistical genetics
- Software version
- 1.94.1 (with updates up to 1.95.4)
- Supported architectures
- x86_64 and ARM64 / AArch64
- Input formats
- PLINK files (.fam, .bim), BGEN (>= v1.2), PGEN, hybrid PGEN, and GWAS summary statistics
What you can use it for
- SNP-based heritability estimation
- Quantify the proportion of phenotypic variance explained by common SNPs using GREML, GREMLd, or Haseman-Elston regression.
- Genetic correlation analysis
- Estimate the genetic correlation between two complex traits or diseases using bivariate GREML or summary statistics.
- Biobank-scale association testing
- Run mixed linear model or generalized linear mixed model GWAS on quantitative and binary traits using fastGWA and fastGWA-GLMM with a sparse GRM.
- Conditional and joint SNP association
- Perform multi-SNP conditional and joint association analyses (COJO and mtCOJO) using GWAS summary statistics.
- Mendelian randomisation
- Test for causal relationships between risk factors and common diseases using generalised summary-data-based Mendelian randomisation (GSMR).
Run it here
GCTA is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
Strengths and limits
Good at
- Supports multiple input data types, including raw genotype matrices, imputed dosage files in BGEN and PGEN formats, and GWAS summary statistics.
- Provides specialized, resource-efficient mixed-model algorithms (fastGWA and fastGWA-GLMM) tailored for large biobank-scale datasets using sparse GRMs.
- Enables detailed heritability partitioning by chromosome, functional annotation, and allele frequency spectrum.
Watch out for
- Compilation on Linux requires specific library configurations and dependencies, including an exact version of Eigen (3.3.7) and Intel MKL or OpenBLAS.
- Windows and macOS versions have not been as comprehensively tested by the developers as the Linux releases.
- Compiling directly on macOS or Windows is discouraged due to compilation complexity, requiring users to rely on precompiled binaries.
Questions
Can I run GCTA online without installing it?
GCTA is preinstalled on the machine behind every research conversation. Ask the agent in plain language, or give it the exact command — it runs it, reads the output and explains it.
What licence is GCTA under?
GCTA is published under its authors' own licence; the project page at https://yanglab.westlake.edu.cn/software/gcta/ states it.
What genotype file formats does GCTA support?
GCTA supports standard PLINK genotype formats (.fam and .bim with genotypes or hybrid PGEN), BGEN format (version 1.2 or higher, including imputed dosages), and PGEN files.
How does fastGWA differ from standard GREML or MLMA?
fastGWA uses a sparse genetic relationship matrix combined with fast variance component estimation to scale linear mixed model association testing to biobank-sized cohorts far faster than dense-matrix MLMA.
Can GCTA test for non-additive genetic effects?
Yes, GCTA includes the GREMLd module to estimate dominance variance among unrelated individuals using GWAS data.
Can GCTA analyze traits using only GWAS summary statistics?
Yes, modules such as COJO, mtCOJO, fastBAT, and GSMR carry out conditional association testing, gene-based tests, and Mendelian randomisation using summary statistics.
How to cite it
Yang et al. (2011) GCTA: a tool for Genome-wide Complex Trait Analysis. Am J Hum Genet. 88(1): 76-82.
Written from GCTA's own README and project page. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.