What are your favorite methods for dissecting pleiotropic relationships, especially directly (with effect size estimates from GWAS across multiple traits) or alternatively, via latent variable approaches? Reviewing this space and hoping not to miss important contributions