The PBAT program (Lange et al. 2003) was employed to determine the association of individual SNPs with each ND measure. Haplotype association was analyzed by the FBAT program (Horvath et al. 2004). P values were evaluated under the null distribution of no linkage and no association. Additive, dominant, and recessive models were tested by incorporating age and gender as covariates for the AA and EA samples; for the pooled sample, age, gender and ethnicity covariates were employed. An association of an individual SNP with ND was considered significant if its P value was less than the corrected P value, based on the method of SNP spectral decomposition (SNPSpD) approach for multiple testing (Nyholt 2004). For haplotype analysis, Bonferroni correction was used to calculate the corrected P-value.
