If you have questions or would like to read the full thesis feel free to reach out anytime.

Always happy to chat science! 🧬🧠

This work highlights the importance of modeling substance use traits across severity levels and including tobacco-related measures🚬 when examining cannabis-related outcomes.🍁

Overlooking these distinctions could obscure true etiological pathways—or lead to spurious associations.🧬

These findings suggest that links between cannabis and certain psychiatric disorders may reflect shared liability with tobacco use, rather than direct effects.

Similarly, genetic overlap between cannabis and other substance use may be better explained by shared liability, not causal progression.

In networks including cocaine and opioid dependence, no significant partial genetic correlations with cannabis or tobacco traits emerged—suggesting no unique genetic link.

This challenges causal “gateway” models and instead points to shared liability through broader traits like externalizing.

Including Major Depressive Disorder (MDD) in the network revealed partial genetic correlations with both cannabis and tobacco traits—highlighting shared liability.

Crucially, MDD was more strongly linked to heavier or disordered use, not mere experimentation.

Using GNA, I modelled genetic relationships among 3 cannabis and 4 tobacco-related traits.

Interestingly, the link between lifetime cannabis use and cannabis use disorder was mediated by frequency of use—mirroring expected patterns:
experimentation → frequent use → disordered use.

Genetic correlations were notably lower for the cannabis-specific factor (vs common) factor with:
Major Depressive Disorder
ADHD
Insomnia
Anxiety
This suggests that previously reported associations with cannabis may be partially explained by shared liability with tobacco use.

I then estimated genetic correlations between external traits and:
(1) a cannabis-specific factor and
(2) a cannabis-common factor.
Lower rGs with the cannabis-specific factor suggest that observed associations may be driven by shared liability with tobacco.

Using GenomicSEM, I built a hierarchical model where (3) cannabis and (4) tobacco phenotypes loaded onto separate latent factors.
Then, through a Cholesky decomposition, I isolated a cannabis-specific latent factor—capturing variance independent of tobacco.

Leveraging the latest GWAS data, I applied Genomic Structural Equation Modelling (GenomicSEM) and Genomic Network Analysis (GNA) to untangle the shared and unique genetic architecture of cannabis use and tobacco use. 🧬📈

Does cannabis cause psychiatric disorders? 🌿🍁🧠

Is it a gateway – or are we missing the bigger picture? 💉

Given its own links to mental illness, could tobacco co-use and shared genetic risk be the hidden confounder in the cannabis-psychopathology link? 🚬