I wrote a little bit about a cool recent paper looking at heritability estimates from very large registry data, and how we still really don't understand why outcomes track in families. A short 🧵: