Evaluation of estimated genetic values and their application to genome-wide investigation of systolic blood pressure

Table 1 Description of GWA results for EGVs and SBP across simulations.

	SBP	EGV_ped	EGV_snp	EGV_seq
Total number of significant (sig) SNPs in 100 simulations	166	134	191	273
Sig SNPs with minor allele frequency (MAF) <0.05	12.0%	5.2%	11.5%	6.6%
Sig SNPs with MAF <0.01	1.2%	0.7%	1.0%	0.7%
Mean number of sig SNPs	11.9	7.8	13.9	14.1
Median number of sig SNPs	8	2	8	8
stdev in number of sig SNPs	15.8	15.0	18.4	19.3
False-discovery rate (FDR)	7.7%	5.7%	6.7%	7.7%
FDR for SNPs not seen in SBP	-	52.1%	50.0%	40.4%
Smaller average p value for sig SNPs	-	30.0%	57.0%	48.0%
Simulations identifying more SNPs than SBP	-	14	39	36
Simulations identifying fewer SNPs than SBP	-	77	18	30
Simulations identifying same number of SNPs as SBP	-	9	43	34

For 100 replicates of simulated SBP, GWA was performed on the raw data, EGV_ped, EGV_snp, and EGV_seq. The following table gives descriptive statistics for significant SNPs (p < 5 × 10^-⁸) by method. The last five rows illustrate the performance of the EGV method relative to the raw SBP GWA.

ISSN: 1753-6561