Multi-ancestry conditional and joint analysis applied to GWAS summary statistics

Abstract

Manc-COJO (Multi-Ancestry Conditional and Joint Analysis) is a C++ software package for identifying independent association signals from multi-ancestry GWAS summary statistics. By jointly modelling association signals across ancestry groups while accounting for ancestry-specific linkage disequilibrium, Manc-COJO improves the detection of independent associations and reduces false positives compared with single-ancestry COJO and existing ad hoc multi-ancestry approaches. This repository contains the Manc-COJO source code, simulation and real-data analysis scripts, and example datasets. The software is available at https://github.com/light156/multi-ancestry-COJO, and the documentation and tutorial are available at https://light156.github.io/multi-ancestry-COJO-docs/.