Not really AM, but more that sibships may have the same genetic relatedness, but different _levels_ of "education-friendly" genes. That between-family difference in genetics will then be soaked up in a0 (i.e. your estimate of c²). Or am I wrong?
Apologies if this is a silly question, but if it's based on SNPs, isn't it a bit similar to SNP heritabilities which are usually pretty far from twin heritabilities? Meaning that the IBD estimate may not be accurate?
Sib-regression is pretty solid and distinct from SNP heritability methods like GREML or LDSC, though I believe it suffers from the same big flaw as twin models in that it can’t very well distinguish GxE, dominance, or epistasis from G
I was referring to "We use the sib pairs’ molecular genetic data to compute their genetic relatedness π. The technical expression for what we refer to as the pair’s genetic relatedness is their “identity by descent” (IBD)". I was wondering if accurate relatedness can be inferred from SNPs alone
Great work! Would you recommend using DF regression in the future over SEM? I'm curious whether the speed/power trade-off is enough to justify the use of the former 👀
Comments
https://academic.oup.com/bioinformatics/article/26/22/2867/228512
(speaking as a fraternal twin).