Table 2.

Naive, replication and bias-reduced genome-wide BR-squared regression coefficient (β) estimates of the significant SNP associated with HbA1c in the DCCT samples

Chr	SNP	Pos (Mb)	MAF	Risk allele	p valuea	Naive betab	Replication betac	BR-squared betad	SDe	Percentage reductionf
10	rs1358030	108.1	0.36	C	4.66E–09	0.045	0.005	0.003	0.013	93%

^aThe association p value using the 667 samples from the DCCT CON group

^bThe naive estimate of the regression coefficient beta using the 667 CON samples without accounting for the threshold effect

^cThe beta estimated from the independent 637 samples from the DCCT INT group

^dThe bias-reduced beta estimate provided by BR-squared using 1,000 level 1 bootstrap samples and with α = 5 × 10⁻⁸

^eSD of the BR-squared beta estimate based on 100 level 2 bootstrap samples

^fThe percentage reduction in the estimates as measured by 1 − bootstrap.beta/naive.beta