Extended Data Fig. 7: Substance use disorder trait loadings onto common genetic factor, AFR ancestry.

The common genetic factor was produced using GenomicSEM-0.0.5c from GWAS of each input trait shown in the diagram. Because African ancestry reference files are not provided with the GenomicSEM-0.0.5c suite, we used LD scores from the Pan-UK Biobank (https://pan.ukbb.broadinstitute.org 2020.) to compute LD matrices and correlations. Analysis was then performed after filtering SNPs to include those with MAF > 0.01 in the 1000 Genomes African ancestry population (https://doi.org/10.1093/nar/gkz836). Values attached to arrows from Fg to trait represent loading of trait onto the common factor. Bottom values represent residual variance in each trait that is unexplained by the common genetic factor. Standard error values are in parentheses.