DIMPLE-GWAS framework identifies 25 latent phenotypes in ~33K European ancestry UK Biobank participants

This cohort study utilized the DIMPLE-GWAS framework applied to brain imaging phenotypes within a population of ~33K European ancestry participants from the UK Biobank. The study compared this approach to input individual imaging phenotypes (IDPs) and prior GWAS of individual IDPs. The primary outcome was the identification of latent genetic architecture across high-dimensional pleiotropic phenotypes. Secondary outcomes included the heritability of latent phenotypes, power for locus discovery, alignment with conventional brain atlas boundaries, and genetic relationships with neurologic, psychiatric, cognitive, and behavioral phenotypes.

The analysis identified 25 biologically interpretable latent phenotypes. These latent phenotypes demonstrated substantially greater heritability than the input IDPs. Additionally, the study identified 104 genomewide-significant loci not reported in prior GWAS of individual IDPs. The resulting structure was validated in the independent ABCD cohort.

Safety data, adverse events, and discontinuations were not reported as this was a genetic analysis study. Key limitations include that the study phase was not reported and follow-up was not reported. Funding or conflicts were not reported. The practice relevance was not reported. Causality was not assessed. These findings are observational and should be interpreted with caution regarding clinical application.