Abstract
Circadian rhythms are daily internal predictive biological cycles. Numerous studies have associated human circadian gene polymorphisms with phenotypic changes and multiple disease risks. Subsequent replications have often been negative in other populations, possibly due to poor linkage disequilibrium (LD) between the reported single nucleotide polymorphisms (SNPs) and underlying phenomena, or differing LD across populations. Alternatively, non-human circadian studies have reported population- specific effects, often expressed as latitudinal dines. Genome-wide genotyping results can produce distinctions. Haplotypes including separate association SNPs can unify multiple results. Strong LD across populations improves the likelihood of population-specific effects. Unusual patterns can possibly predict undiscovered effects. The importance of in silico techniques increases as ongoing projects seek to sequence all common SNPs. The present study has examined the variation patterns of 35 upstream, downstream, and core clock genes. 20 genes displayed unusual results using multiple screening measures. A possible 7-gene African-population haplogroup network was identified. While selection across the class of circadian genes was ruled out, CLOCK and FBXW11 are candidates for balancing and positive selection, supported by multiple genome- wide frequency spectrum distributions. CLOCK was examined in a UK-based sample in the most dense diurnal preference candidate gene study yet performed. An eveningness haplotype independent of the previously reported 3111C allele was identified. This haplotype and 3111C combined were important for intermediate preference. Strong LD, including 3 haplogroups and a 122-SNP 3111C haplotype shared by all 3 HapMap populations, united distinct association study results into population-specific phenotypes unlikely to be due to differing or poor LD across populations. The eveningness association was age-dependent. A non-linear relationship between ageing and diurnal preference was also uncovered, particularly for morning functioning and evening sleepiness. A possible circadian-climate interaction was examined. 14 genes had strongly impacted LD patterns. Near-exact relationships between climate (but not latitude) SNPs and diurnal preference were found in PERI and PER2, while an extrapolated relationship affected PER3. Strong relationships demonstrated negative or even opposite results in other populations. Combined, these results suggest that human circadian phenotypes, and possibly related disease risks, are likely altered by the environment. This raises the importance of future multiple-population studies, but predicts increased complexity in their interpretation.