Genetic basis of immune-mediated necrotizing myopathy: A genomic structural equation modeling and AI-driven...

Immune-mediated necrotizing myopathy (IMNM) constitutes an autoimmune myopathy precipitated by autoantibodies, marked by rapid progression and recurrent exacerbations, with no approved curative therapies currently available. Genetic determinants serve as pivotal etiological contributors; however, comprehensive genome-wide association study (GWAS) data elucidating IMNM's genetic architecture remain absent.

We implemented Genomic Structural Equation Modeling (Genomic SEM) to perform a large-scale multivariate GWAS, probing shared genetic architectures across traits to estimate single nucleotide polymorphisms (SNPs) associated with independent variant associations. Multiple transcriptome-wide association analyses were subsequently deployed to identify potential mediation loci with high IMNM correlation, spanning tissue stratifications, cellular hierarchies, and genomic annotations. AI modeling coupled with protein molecular dynamics simulations was then employed to decipher how such loci modulate protein structure and functional dynamics.

This study integrated GWAS data from six phenotypes through a multi-scale approach and systematically analyzed 5049,276 common variants within the Genomic SEM framework. Using FUMA for functional annotation and independent filtering, we identified 128 statistically significant lead loci. These loci may mediate the genetic risk of IMNM through latent factor effects. Protein modeling and dynamics simulations demonstrated that mutations undermine protein structural stability and disrupt the poised balance of functional domain flexibility, thereby compromising functional protein expression.

Our study pioneers a comprehensive genetic framework for IMNM by applying GWAS to an indirectly assessed phenotype, articulating a panoramic view of IMNM's hereditary landscape and the molecular mechanisms through which genetic variations govern disease pathology.