Haplotypes of common SNPs can explain missing heritability of complex diseases. Seems to imply local epistasis.http://biorxiv.org/content/early/2015/07/12/022418 …
-
If we assume epistasis is mostly local, this dramatically shrinks the number of GxG interactions to test in GWAS LASSO.
But the high LD related correlations imply large standard errors which further implies we still need v. large samples. Sigh.
