Response: genetic admixture in sickle cell disease

Genetic studies require careful thought about ancestry in the study design and analysis to avoid population stratification bias and to maximize the power of finding novel variants.¹  If both the genetic marker and the phenotype vary with respect to ancestry, then a spurious association will occur between the genetic marker and phenotype if one does not adjust properly for ancestry. In this issue of Blood, Silva et al examine the level of admixture in a cohort of Brazilian sickle cell patients and find that patients with sickle cell disease have a wide range of African admixture (15%-85%).²  They correctly acknowledge that one must appropriately adjust for admixture to avoid false positive findings and that this population may be useful for admixture mapping. However, admixture mapping is only useful if the phenotype of interest also varies with admixture and further research is required to establish this relation in this population. In a cohort of African Americans with sickle cell disease we did not find a significant association between admixture (measured by the first principal component from a principal component analysis) and fetal hemoglobin.³  However, one should note that the African Americans in this study on average did not have high levels of Caucasian admixture.⁴  Admixture mapping has been successfully used to find novel genetic variants and regions for other phenotypes that were related to admixture such as white cell counts and prostate cancer.^5,6  Furthermore, examining ancestry in genetic studies of sickle cell disease could lead to novel loci that are either more prevalent or specific to certain ethnic groups. For example, sickle cell patients from the Southwestern Province of Saudi Arabia have fetal hemoglobin (HbF) levels twice as high as African Americans despite having similar HBB haplotypes.⁷  Furthermore, sickle cell patients from the Eastern Province have even higher levels of HbF (mean [SD] 30.4 ± 6.9).⁸  These findings suggest that there are HbF-associated variants that are more prevalent or specific to the Saudi population and that leveraging on ancestry in the genetic analysis can help identify novel variants.