mFABIO method extends TWAS fine-mapping to binary traits across multiple tissues

Researchers at the University of Michigan, led by Haihan Zhang, Kevin He, Lam C. Tsoi, and Xiang Zhou, have published mFABIO in PLOS Genetics — a multi-tissue transcriptome-wide association study (TWAS) fine-mapping method designed specifically for binary traits.

Existing TWAS fine-mapping approaches jointly model multiple genes to prioritise candidates for causal roles at a given locus, but they have largely been developed for quantitative traits and typically draw on gene expression data from a single tissue. mFABIO addresses both limitations. It uses a probit model to link genetically regulated expression (GReX) of genes across multiple tissues simultaneously to binary trait outcomes, enabling joint causal gene and tissue prioritisation within a single statistical framework.

The method is relevant to the growing body of GWAS loci for common diseases where the causal gene and the tissue in which its dysregulation matters remain uncertain. By modelling binary outcomes directly rather than approximating them as quantitative, mFABIO avoids the well-known calibration problems that arise when linear models are applied to case-control data at the TWAS stage.

The paper sits within the active field of post-GWAS functional inference, complementing recent work on colocalisation and Mendelian randomisation frameworks. Researchers working on common complex diseases — particularly where multi-tissue eQTL reference panels such as GTEx are already in use — may find mFABIO a useful addition to their analytical toolkit.