Bipolar patients display stoichiometric imbalance of gene expression in post-mortem brain samples

Asbjørn Holmgren,Ibrahim Akkouh,Kevin Sean O'Connell,Jordi Requena Osete,Pål Marius Bjørnstad,Srdjan Djurovic,Timothy Hughes

DOI: https://doi.org/10.1038/s41380-023-02398-0

IF: 11

2024-02-13

Molecular Psychiatry

Abstract:Bipolar disorder is a severe neuro-psychiatric condition where genome-wide association and sequencing studies have pointed to dysregulated gene expression as likely to be causal. We observed strong correlation in expression between GWAS-associated genes and hypothesised that healthy function depends on balance in the relative expression levels of the associated genes and that patients display stoichiometric imbalance. We developed a method for quantifying stoichiometric imbalance and used this to predict each sample's diagnosis probability in four cortical brain RNAseq datasets. The percentage of phenotypic variance on the liability-scale explained by these probabilities ranged from 10.0 to 17.4% (AUC: 69.4–76.4%) which is a multiple of the classification performance achieved using absolute expression levels or GWAS-based polygenic risk scores. Most patients display stoichiometric imbalance in three to ten genes, suggesting that dysregulation of only a small fraction of associated genes can trigger the disorder, with the identity of these genes varying between individuals.

biochemistry & molecular biology,neurosciences,psychiatry