Paul McKeigue

Problems of reporting genetic associations with complex outcomes

Inability to replicate many results has led to increasing scepticism about the value of simple association study designs for detection of genetic variants contributing to common complex traits. Much attention has been drawn to the problems that might, in theory, bedevil this approach, including confounding from population structure, misclassification of outcome, and allelic heterogeneity. Other researchers have argued that absence of replication may indicate true heterogeneity in gene-disease associations. We suggest that the most important factors underlying inability to replicate these associations are publication bias, failure to attribute results to chance, and inadequate sample sizes, problems that are all rectifiable. Without changes to present practice, we risk wastage of scientific effort and rejection of a potentially useful research strategy.

SLC24A5, a putative cation exchanger, affects pigmentation in zebrafish and humans.

Lighter variations of pigmentation in humans are associated with diminished number, size, and density of melanosomes, the pigmented organelles of melanocytes. Here we show that zebrafish golden mutants share these melanosomal changes and that golden encodes a putative cation exchanger slc24a5 (nckx5) that localizes to an intracellular membrane, likely the melanosome or its precursor. The human ortholog is highly similar in sequence and functional in zebrafish. The evolutionarily conserved ancestral allele of a human coding polymorphism predominates in African and East Asian populations. In contrast, the variant allele is nearly fixed in European populations, is associated with a substantial reduction in regional heterozygosity, and correlates with lighter skin pigmentation in admixed populations, suggesting a key role for the SLC24A5 gene in human pigmentation.

Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15 000 Swedish men and women born 1915-29.

Abstract Objective: To establish whether fetal growth rate (as distinct from size at birth) is associated with mortality from ischaemic heart disease. Design: Cohort study based on uniquely detailed obstetric records with 97% follow up over the entire life course and linkage to census data in adult life. Subjects: All 14 611 babies delivered at the Uppsala Academic Hospital, Sweden, during 1915-29 followed up to end of 1995. Main outcome measures: Mortality from ischaemic heart disease and other causes. Results: Cardiovascular disease showed an inverse association with birth weight for both men and women, although this was significant only for men. In men a 1000 g increase in birth weight was associated with a proportional reduction in the rate of ischaemic heart disease of 0.77 (95% confidence interval 0.67 to 0.90). Adjustment for socioeconomic circumstances at birth and in adult life led to slight attenuation of this effect. Relative to the lowest fourth of birth weight for gestational age, mortality from ischaemic heart disease in men in the second, third, and fourth fourths was 0.81 (0.66 to 0.98), 0.63 (0.50 to 0.78), and 0.67 (0.54 to 0.82), respectively. The inclusion of birth weight per se and birth weight for gestational age in the same model strengthened the association with birth weight for gestational age but removed the association with birth weight. Conclusions: This study provides by far the most persuasive evidence of a real association between size at birth and mortality from ischaemic heart disease in men, which cannot be explained by methodological artefact or socioeconomic confounding. It strongly suggests that it is variation in fetal growth rate rather than size at birth that is aetiologically important.

Relation of size at birth to non-insulin dependent diabetes and insulin concentrations in men aged 50-60 years

Abstract Objective: To establish whether the relation between size at birth and non-insulin dependent diabetes is mediated through impaired β cell function or insulin resistance. Design: Cohort study. Setting: Uppsala, Sweden. Subjects: 1333 men whose birth records were traced from a cohort of 2322 men born during 1920-4 and resident in Uppsala in 1970. Main outcome measures: Intravenous glucose tolerance test at age 50 years and non-insulin dependent diabetes at age 60 years. Results: There was a weak inverse correlation (r=-0.07, P=0.03) between ponderal index at birth and 60 minute insulin concentrations in the intravenous glucose tolerance test at age 50 years. This association was stronger (r=-0.19, P=0.001) in the highest third of the distribution of body mass index than in the other two thirds (P=0.01 for the interaction between ponderal index and body mass index). Prevalence of diabetes at age 60 years was 8% in men whose birth weight was less than 3250 g compared with 5% in men with birth weight 3250 g or more (P=0.08; 95% confidence interval for difference −0.3% to 6.8%). There was a stronger association between diabetes and ponderal index: prevalence of diabetes was 12% in the lowest fifth of ponderal index compared with 4% in the other four fifths (P=0.001; 3.0% to 12.6%). Conclusion: These results confirm that reduced fetal growth is associated with increased risk of diabetes and suggest a specific association with thinness at birth. This relation seems to be mediated through insulin resistance rather than through impaired β cell function and to depend on an interaction with obesity in adult life.

Cardiovascular disease in women with polycystic ovary syndrome at long‐term follow‐up: a retrospective cohort study

Polycystic ovary syndrome (PCOS) is associated with higher prevalence of cardiovascular risk factors but the relative prevalence of cardiovascular disease in women with PCOS has not previously been reported. We have compared cardiovascular mortality and morbidity in middle‐aged women previously diagnosed with PCOS and age‐matched control women.

SLC2A9 is a newly identified urate transporter influencing serum urate concentration, urate excretion and gout

Uric acid is the end product of purine metabolism in humans and great apes, which have lost hepatic uricase activity, leading to uniquely high serum uric acid concentrations (200–500 μM) compared with other mammals (3–120 μM). About 70% of daily urate disposal occurs via the kidneys, and in 5–25% of the human population, impaired renal excretion leads to hyperuricemia. About 10% of people with hyperuricemia develop gout, an inflammatory arthritis that results from deposition of monosodium urate crystals in the joint. We have identified genetic variants within a transporter gene, SLC2A9, that explain 1.7–5.3% of the variance in serum uric acid concentrations, following a genome-wide association scan in a Croatian population sample. SLC2A9 variants were also associated with low fractional excretion of uric acid and/or gout in UK, Croatian and German population samples. SLC2A9 is a known fructose transporter, and we now show that it has strong uric acid transport activity in Xenopus laevis oocytes.

Control of Confounding of Genetic Associations in Stratified Populations

To control for hidden population stratification in genetic-association studies, statistical methods that use marker genotype data to infer population structure have been proposed as a possible alternative to family-based designs. In principle, it is possible to infer population structure from associations between marker loci and from associations of markers with the trait, even when no information about the demographic background of the population is available. In a model in which the total population is formed by admixture between two or more subpopulations, confounding can be estimated and controlled. Current implementations of this approach have limitations, the most serious of which is that they do not allow for uncertainty in estimations of individual admixture proportions or for lack of identifiability of subpopulations in the model. We describe methods that overcome these limitations by a combination of Bayesian and classical approaches, and we demonstrate the methods by using data from three admixed populations--African American, African Caribbean, and Hispanic American--in which there is extreme confounding of trait-genotype associations because the trait under study (skin pigmentation) varies with admixture proportions. In these data sets, as many as one-third of marker loci show crude associations with the trait. Control for confounding by population stratification eliminates these associations, except at loci that are linked to candidate genes for the trait. With only 32 markers informative for ancestry, the efficiency of the analysis is 70%. These methods can deal with both confounding and selection bias in genetic-association studies, making family-based designs unnecessary.

Epidemiological methods for studying genes and environmental factors in complex diseases

Exploration of the human genome presents new challenges and opportunities for epidemiological research. Although the case-control design is quicker and cheaper for study of associations between genotype and risk of disease than the cohort design, cohort studies have been recommended because they can be used to study gene-environment interactions. Although the scientific relevance of statistical interaction is pertinent, the main disadvantage of the case-control design-susceptibility to bias when estimating effects of exposures that are measured retrospectively-does not necessarily apply when studying statistical interaction between genotype and environmental exposure. Because correctly designed genetic association studies are equivalent to randomised comparisons between genotypes, conclusions about cause can be drawn from genetic associations even when the risk ratio is modest. For adequate statistical power to detect such modest risk ratios, the case-control design is more feasible than the cohort design.

Mortality of Women with Polycystic Ovary Syndrome at Long-term Follow-up

Metabolic disturbances associated with insulin resistance are present in most women with polycystic ovary syndrome. This has led to suggestions that women with polycystic ovary syndrome may be at increased risk of cardiovascular disease in later life. We undertook a long-term follow-up study to test whether cardiovascular mortality is increased in these women. A total of 786 women diagnosed with polycystic ovary syndrome in the United Kingdom between 1930 and 1979 were traced from hospital records and followed for an average of 30 years. Standardized mortality ratios (SMRs) were calculated to compare the death rates of these women with national rates. The SMR for all causes was 0.90 (95% CI, 0.69-1.17), based on 59 deaths. There were 15 deaths from circulatory disease, yielding an SMR of 0.83 (95% CI, 0.46-1.37). Of these 15 deaths, 13 were from ischemic heart disease (SMR 1.40; 95% CI, 0.75-2.40) and two were from other circulatory disease (SMR 0.23; 95% CI, 0.03-0.85). There were six deaths from diabetes mellitus as underlying or contributory cause, compared with 1.7 expected (odds ratio 3.6; 95% CI, 1.5-8.4). Breast cancer was the commonest cause of death (SMR 1.48 based on 13 deaths; 95% CI, 0.79-2.54). We conclude that women with polycystic ovary syndrome do not have markedly higher than average mortality from circulatory disease, even though the condition is strongly associated with diabetes, lipid abnormalities, and other cardiovascular risk factors. The characteristic endocrine profile of women with polycystic ovary syndrome may protect against circulatory disease in this condition.

Skin pigmentation, biogeographical ancestry and admixture mapping

Ancestry informative markers (AIMs) are genetic loci showing alleles with large frequency differences between populations. AIMs can be used to estimate biogeographical ancestry at the level of the population, subgroup (e.g. cases and controls) and individual. Ancestry estimates at both the subgroup and individual level can be directly instructive regarding the genetics of the phenotypes that differ qualitatively or in frequency between populations. These estimates can provide a compelling foundation for the use of admixture mapping (AM) methods to identify the genes underlying these traits. We present details of a panel of 34 AIMs and demonstrate how such studies can proceed, by using skin pigmentation as a model phenotype. We have genotyped these markers in two population samples with primarily African ancestry, viz. African Americans from Washington D.C. and an African Caribbean sample from Britain, and in a sample of European Americans from Pennsylvania. In the two African population samples, we observed significant correlations between estimates of individual ancestry and skin pigmentation as measured by reflectometry (R2=0.21, P<0.0001 for the African-American sample and R2=0.16, P<0.0001 for the British African-Caribbean sample). These correlations confirm the validity of the ancestry estimates and also indicate the high level of population structure related to admixture, a level that characterizes these populations and that is detectable by using other tests to identify genetic structure. We have also applied two methods of admixture mapping to test for the effects of three candidate genes (TYR, OCA2, MC1R) on pigmentation. We show that TYR and OCA2 have measurable effects on skin pigmentation differences between the west African and west European parental populations. This work indicates that it is possible to estimate the individual ancestry of a person based on DNA analysis with a reasonable number of well-defined genetic markers. The implications and applications of ancestry estimates in biomedical research are discussed.