Abstracts

Genetic association studies have had an inconsistent record so far (Colhoun et al. 2003). A single nucleotide polymorphism (G1465A) in the GABA(B) receptor 1 gene has been reported to be associated with non-lesional temporal lobe epilepsy (TLE) in a study of 141 subjects (Gambardella et al. 2003). Subjects with the A allele were at higher risk of TLE. Population admixture was seen in this association study, thus cryptic population stratification remains a possibility.
As population stratification can cause spurious associations, independent replication of association studies is important to verify or refute published associations. We therefore performed a case-control replication study to confirm this association. We recruited 200 subjects with non-lesional TLE using identical case definitions as the first study; 88 healthy subjects were used as controls. DNA was extracted with standard methods, and presence of the G1465A polymorphism assessed using restriction fragment length polymorphisms. Differences in allele frequencies between groups were compared using the chi-square test; reported p-values are two-sided. Our results showed that the A allele was absent in cases (0%, 95% CI 0-2.3%) and present in 0.8% of controls (95% CI 0.2-3.6%); this was not significantly different (p=0.55).
We were unable to replicate the association despite a larger sample size and adherence to existing guidelines for genetic association studies. We suggest that the initial association may be due to uncorrected cryptic population stratification. Though the first study used logistic regression to minimise overt stratification, this does not minimise cryptic stratification, which can bias even well-designed studies(Freedman et al. 2004). Other possible reasons for non-replication such as inadequate sample size or genotyping errors are less likely. We therefore suggest that new methods for correcting cryptic population stratification, such as genomic control, should be used routinely to verify positive findings in genetic association studies. Independent replication of genetic association studies remains an important method of verifying published associations, and cautious interpretation of association studies is emphasised.
References
Colhoun, HM. et al. (2003). Problems of reporting genetic associations with complex outcomes. Lancet 361: 865-72.
Freedman, ML. et al. (2004). Assessing the impact of population stratification on genetic association studies. Nat Genet 36(4): 388-93.
Gambardella, A, et al. (2003). GABA(B) receptor 1 polymorphism (G1465A) is associated with temporal lobe epilepsy. Neurology 60(4): 560-3. (Supported by NHMRC Australia [amp] by Bionomics Ltd. NCKT is supported by research fellowships from NHG and NMRC, Singapore.)