RESEARCH AREAS
Genetics of substance use disorders
I use genome-wide methods to better understand the genetic architecture of substance use disorders. These projects require large-scale collaborations via international consortia to achieve adequate statistical power. The resulting summary statistics can be used in many downstream analyses (e.g., creating polygenic scores). In 2020, we published the largest genome-wide association study of cannabis use disorder to date in collaboration with researchers in Denmark and Iceland, identifying two genomic risk loci associated with cannabis use disorder. Interestingly, we saw that cannabis use disorder showed divergent correlations with several traits of interest, including body mass index and educational attainment, when compared with cannabis ever-use (a measure of whether someone has ever tried cannabis) - while cannabis use disorder is correlated with greater BMI and lower educational attainment, cannabis ever-use is correlated with lower BMI and greater educational attainment. This suggests that the genetic factors underlying problematic cannabis use at least partly diverge from those of "typical" use, or ever trying the substance.
Unique and shared genetic factors underlying co-occurring mental health conditions and behaviors
Using methods such as genomic structural equation modeling, I examine the genetic overlap (and points of divergence) among different behaviors and mental health conditions. For example, a recent study from our lab identified relatively large genetic correlations between suicide-related behaviors and substance use disorders; furthermore, these relationships persist even when accounting for the genetics of depression, a risk factor often evaluated in clinical settings. These findings underscore the transdiagnostic nature of suicide-related behaviors.
Autozygosity and complex traits & behaviors
I am interested in using genome-wide data to ask evolutionarily-informed questions about human behaviors and psychiatric disorders - for example, do we see evidence that genetic risk variants for schizophrenia have been selected against over evolutionary time? This type of analysis can reveal the extent to which different types of genetic variants may contribute to risk of a certain disorder (e.g., rare, recessive variants). Recent work in the lab investigated levels of autozygosity (when alleles are identical by descent) in a contemporary, American sample, looking at the association between autozygosity and cognitive performance. Interestingly, we found much lower levels of autozygosity in this relatively young sample than in previously-studied samples of older adults, suggesting that overall levels of autozygosity might be decreasing over generations. This was confirmed in a meta-analysis of the UK Biobank, Million Veteran Program, and All of Us that was recently published. Future directions include looking at associations with autozygosity in between-sibling and within-sibling designs, as within-sibling comparisons should avoid potential confounding that can plague population-level estimates.