Evaluating the estimation of genetic correlation and heritability using summary statistics

Area of Science:

• Population Genetics
• Statistical Genomics

Background:

• Novel statistical methods are emerging to quantify shared heritability across distinct ancestral populations.
• The Popcorn method (Brown et al. 2016) estimates genetic correlation using summary statistics.

Purpose of the Study:

• To thoroughly evaluate the Popcorn method's performance under various parameters.
• To determine optimal conditions for accurate estimation of genetic correlation and heritability.

Main Methods:

• The Popcorn method was assessed across different sample sizes, SNP counts, and reference panel scenarios.
• Performance was evaluated using various population pairs, including homogeneous and admixed groups.
• The impact of inappropriate reference panels was specifically investigated.

Main Results:

• Minimum sample sizes for reference panels, summary statistics, and SNP counts were identified for accurate estimation.
• Required sample size and SNP number for stable estimates showed direct proportionality with heritability.
• Misrepresenting the reference panel led to overestimations of genetic correlation (20%) and heritability (60%).
• Application to homogeneous (EUR) and admixed (ASW) populations resulted in underestimation of genetic correlation (15%).

Conclusions:

• Accurate estimation of genetic correlation and heritability using Popcorn requires careful consideration of sample size, SNP count, and reference panel generalizability.
• Caution is advised when applying these methods to ensure robust and reliable etiologic insights from diverse populations.