Psychiatric genome-wide association study enrichment shows promise for future psychopharmaceutical discoveries

Abstract

Background: Innovation in psychiatric therapeutics has stagnated on known mechanisms. Psychiatric genome-wide association studies (GWAS) have identified hundreds of genome-wide significant (GWS) loci that have rapidly advanced our understanding of disease etiology. However, whether these results can be leveraged to improve clinical treatment for specific psychiatric disorders remains poorly understood.

Methods: In this proof-of-principal evaluation of GWAS clinical utility, we test whether the targets of drugs used to treat Attention Deficit Hyperactivity Disorder (ADHD), Bipolar Disorder (BiP), Generalized Anxiety Disorder (GAD), Major Depressive Disorder (MDD), Post-Traumatic Stress Disorder (PTSD), Schizophrenia (SCZ), Substance Use Disorders (SUDs), and insomnia (INS), are enriched for GWAS meta-analysis findings.

Results: The genes coding for treatment targets of medications used to SCZ, BiP, MDD, and SUDs (but not ADHD, PTSD, GAD, or INSOM) are enriched for GWS loci identified in their respective GWAS (ORs: 2.78-27.63; all ps <1.15e-3). Enrichment is largely driven by the presence of a GWS locus or loci within a gene coding for a drug target (i.e., proximity matching). Broadly, additional annotation (i.e., functional: Combined Annotation Dependent Depletion [CADD] scores, regulomeDB scores, eQTL, chromatin loop, and gene region; statistical: effect size of genome-wide significant SNPs; Z-score of SNPs; number of drug targets implicated by GWAS), with the exception of weighting by the largest SNP effect size, does not further improve enrichment across disorders. Evaluation of prior smaller GWAS reveal that more recent larger GWAS improve enrichment.

Conclusions: GWAS results may assist in the prioritization of medications for future psychopharmaceutical research.

Fig. 1. GWAS implicates current treatments for psychiatric disorders.

A Odds ratio of a drug implicated by GWAS being a known treatment for a psychiatric disorder, controlling for the number of drug targets of each drug. Those results that were significant after Bonferroni are plotted with their standard error. MDD major depressive disorder, ADHD attention deficit hyperactivity disorder, GAD generalized anxiety disorder, PTSD post-traumatic stress disorder. B The link between a treatment for that disease and the drug targets from GWAS is shown as a Sankey Network Diagram. Larger colored edges signify genes (left) and drugs (right) that are more connected. These are the positive enrichment cases that are driving our significant findings.

Fig. 2. Enrichment for various genomic annotations.

For each disorder we plot the enrichment when genomic annotations are used instead of 0 or 1 any-gene pairing. Each dot is the beta and CI’s for that beta. Abs absolute value or that the absolute value of SNP effect sizes was used, CADD combined annotation dependent depletion, pLI loss intolerance probability score. When points are missing from the plot, it is because no Exonic variant was a drug target after pruning SNPs for linkage disequilibrium. MDD major depressive disorder, SUD Substance use disorders.