Authors' reply

We thank Yeebo for this comment on our paper, raising an important point that the frequency of genetic polymorphisms tends to vary between different ethnic groups. In our study, we had sought to investigate the association of the BDNF Val66Met polymorphism with imaging phenotypes in the context of major depressive disorder, rather than an investigation of an association for the polymorphism with major depressive disorder. As Yeebo states that we had sought ‘to determine whether the same neural effects would be observed in healthy individuals’, we would like to clarify that we had investigated whether the effects of the polymorphism on specific brain regions in healthy individuals would also be observed in the same brain regions in individuals with major depressive disorder.

Yeebo's general comment on the differing frequency of the genetic polymorphisms between different ethnic groups has long been a source of difficulty in allelic association studies because – unless there is careful ethnic matching of cases and controls – spurious results can arise, including both false-positive and false-negative findings. For this reason, we had restricted our study to individuals of White European origin.

Yeebo draws attention to the possibility that the association between the BDNF Val66Met polymorphism and major depressive disorder is specific to certain ethnic groups, and this raises another but slightly different interesting issue. Disease associations with polymorphisms arise either because the polymorphism itself plays a causal role or because it is in linkage disequilibrium with another variant that has a causal role. The degree of linkage disequilibrium between adjacent polymorphisms is in general approximately proportional to their distance apart but is also dependent on other factors including allelic frequency, so ethnic differences in observed linkage disequilibrium may arise. It has usually been assumed that because the BDNF Val66Met polymorphism is functional, its role in depression is likely to be causal. If the association between this polymorphism and major depressive disorder were found to vary between different ethnic groups, this would be attributable to either different Met allele frequencies between the ethnic groups or the Val66Met polymorphism being in linkage disequilibrium with another variant in BDNF that has a causal role in major depressive disorder. ¹

There are also publicly available data on the Single Nucleotide Polymorphism Database (www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs=6265) which indicate that in African and African American populations, the frequency of the Met allele seems to be rare (e.g. Yoruba, Met frequency 0.0004 in a group of 226 sampled; African Americans 0.05 in a group of 90) so that a genetic association analysis with depression would require very large samples.

Although we would anticipate the functional impact of being a Met carrier to be similar across population groups, we agree with Yeebo that this would warrant further investigation.