Shared genetic variants across substance use disorders implicate common neurobiological pathways, a genome-wide mixed...

Substance use disorders (SUDs) are highly heritable, but the extent of shared and distinct genetic architecture across different SUDs is unclear.

To compare the genetic architectures of alcohol use disorder (AUD), cannabis use disorder (CUD) and opioid use disorder (OUD) and to identify shared and unique genetic loci.

We analysed large-scale genome-wide association study (GWAS) summary statistics from individuals of European ancestry recruited in Europe and the USA. The mixture model MiXeR was used to estimate the unique genetic architecture characteristics of each SUD, including its polygenicity, single nucleotide polymorphism (SNP)-heritability and discoverability, a measure of the distribution of genetic signal across all causal variants. Pairwise conditional/conjunctional false discovery rate (cond/conjFDR) analyses identified shared loci, followed by biological annotation of implicated genes.

AUD demonstrated the highest polygenicity, followed by CUD and OUD. SNP-based heritability was 0.10 for AUD and OUD and 0.01 for CUD. Discoverability was highest for OUD, followed by AUD and CUD. Currently, genome-wide significant SNPs explain 2.0% of AUD, 0.3% of CUD and 0.2% of OUD variance. Cond/conjFDR identified 39 novel loci for AUD, 10 for CUD and 1 for OUD. Of implicated genes, most were expressed in the brain, including several involved in gamma-aminobutyric acid and dopaminergic neurotransmission, opioid neurophysiology, myelination, DNA recombination, apoptosis and ubiquitin-dependent protein catabolism.

SUDs have polygenic architectures with many shared loci and are similar with regards to some characteristics. However, the level of polygenicity differs across SUDs, with AUD being considerably more polygenic than OUD, whereas CUD is intermediate in terms of its polygenicity. The novel loci implicate genes primarily expressed in the brain, involving a variety of biological functions. The findings expand our view of the aetiology of these disorders, while supporting the hypothesis of a shared set of pleiotropic SUD genes.