Chunk #27 — ONLINE METHODS — Regression Weights

Source

LD Score regression distinguishes confounding from polygenicity in genome-wide association studies.

Embedded

yes

Text

In order to produce an efficient regression estimator, we must deal with two problems. First, χ2-statistics at SNPs in LD are correlated. Second, the χ2-statistics of variants with high LD Score have higher variance than the χ2-statistics of variants with low LD Score (heteroskedasticity).