None of 10 Previously Reported Cannabis Genes Replicated in a Large Twin Study | Research

The takeaway

A study of 7,452 twins tested 10 candidate genes previously linked to cannabis use and found that none reached even nominal statistical significance, suggesting earlier findings may have been false positives.

Read this if you are interested in the genetics of cannabis use and why specific gene findings often fail to replicate.

0 of 10 candidate genes replicated at even p<0.05

What the researchers found

Researchers used a large Australian twin family sample to test whether 10 genes previously reported to be associated with cannabis use actually replicated. Using a gene-based association test (which is more powerful than single-variant tests), none of the candidate genes reached even the most lenient significance threshold (p<0.05).

The authors concluded that the failure to replicate likely reflects limited understanding of cannabis neurobiology, potential publication bias in earlier studies, and the likelihood of false-positive findings in the original reports.

Why it matters

While cannabis use is clearly heritable (twin studies show this consistently), identifying the specific genes involved has proven extremely difficult. This large non-replication study serves as a cautionary example about the reliability of candidate gene findings in psychiatric genetics.

The numbers in context

7,452 participants. 10 candidate genes tested. None reached p<0.05. This is consistent with the broader "replication crisis" in candidate gene studies for behavioral traits.

How the study worked

Gene-based association testing in 7,452 participants from an Australian twin family sample. Tested 10 candidate genes previously reported as associated with lifetime frequency of cannabis use. The gene-based approach aggregates signal across all variants within each gene, providing more power than single-SNP tests.

What this study cannot tell us

The study tested lifetime frequency of cannabis use, not cannabis dependence, which may have different genetic architecture. The Australian sample may differ genetically from the populations in which the original associations were found. Gene-based tests, while more powerful for some signals, may miss specific variants.

How to read the evidence

Well-powered replication study in a large twin sample; provides strong evidence against the tested candidate genes.

When this study was published

Published in 2012. Genome-wide association studies have since identified some genetic variants associated with cannabis use, though effect sizes are very small.

The bigger picture

This study appeared at a pivotal moment in psychiatric genetics, as the field was shifting from small candidate gene studies to large genome-wide association studies (GWAS). The failure of all 10 candidates to replicate in a well-powered sample illustrated why this shift was necessary.

Questions still open

Which, if any, specific genes genuinely influence cannabis use? Will genome-wide approaches succeed where candidate gene studies have failed? How much of cannabis use heritability can eventually be explained by identified genetic variants?

Common questions

Is cannabis use genetic?

Twin studies consistently show that cannabis use is heritable, with genetics explaining roughly 40-60% of the variation. However, identifying the specific genes responsible has proven very difficult. This study found that 10 genes previously claimed to be involved did not replicate, suggesting the earlier findings were false leads.

Why do genetic findings for cannabis use fail to replicate?

Several factors contribute: the original studies were often too small, leading to statistical flukes being mistaken for real findings. Publication bias means positive results get published while failed replications do not. And cannabis use is likely influenced by many genes, each with tiny effects, making individual gene identification extremely challenging.

Read the original research

No association of candidate genes with cannabis use in a large sample of Australian twin families.

Addiction biology, 17(3), 687-90

View on PubMed ↗Publisher’s paper ↗

Citation

Verweij, Karin J H; Zietsch, Brendan P; Liu, Jimmy Z; Medland, Sarah E; Lynskey, Michael T; Madden, Pamela A F; Agrawal, Arpana; Montgomery, Grant W; Heath, Andrew C; Martin, Nicholas G. (2012). No association of candidate genes with cannabis use in a large sample of Australian twin families.. Addiction biology, 17(3), 687-90. https://doi.org/10.1111/j.1369-1600.2011.00320.x