Multiomic prioritisation of risk genes for anorexia nervosa.

BACKGROUND

Anorexia nervosa (AN) is a psychiatric disorder associated with marked morbidity. Whilst AN genetic studies could identify novel treatment targets, integration of functional genomics data, including transcriptomics and proteomics, would assist to disentangle correlated signals and reveal causally associated genes.

METHODS

We used models of genetically imputed expression and splicing from 14 tissues, leveraging mRNA, protein, and mRNA alternative splicing weights to identify genes, proteins, and transcripts, respectively, associated with AN risk. This was accomplished through transcriptome, proteome, and spliceosome-wide association studies, followed by conditional analysis and finemapping to prioritise candidate causal genes.

RESULTS

We uncovered 134 genes for which genetically predicted mRNA expression was associated with AN after multiple-testing correction, as well as four proteins and 16 alternatively spliced transcripts. Conditional analysis of these significantly associated genes on other proximal association signals resulted in 97 genes independently associated with AN. Moreover, probabilistic finemapping further refined these associations and prioritised putative causal genes. The gene , for which increased genetically predicted mRNA expression was correlated with AN, was strongly supported by both conditional analyses and finemapping. Pathway analysis of genes revealed by finemapping identified the pathway (overlapping genes = , , , ) as statistically overrepresented.

CONCLUSIONS

We leveraged multiomic datasets to genetically prioritise novel risk genes for AN. Multiple-lines of evidence support that is associated with AN, whilst other prioritised genes were enriched within immune related pathways, further supporting the role of the immune system in AN.