Larry D. Atwood, Fellow

Complex Genetic Analysis of Blood Pressure

  A major problem in the genetics of blood pressure is the detection and location of the genes that cause hypertension (high blood pressure). This is a daunting task, as it is widely believed that blood pressure is under the control of multiple genes and that these genes interact with each other and with the environment in complex ways. Current methods are either incapable of modeling this complexity or, if they do have the theoretical capability, they require large databases and are computationally demanding.

  These researchers had a large dataset at their disposal: the San Antonio Family Heart Study is composed of approximately 1000 Mexican Americans organized in 25 large families. Each individual has undergone an extensive data gathering protocol in a clinical setting. An extensive database of blood pressure and other variables on diet, exercise, and medication was available to the researchers. Each individual was genotyped for 400 highly polymorphic markers,0 evenly distributed across the 22 chromosomes.

  Using this dataset, the researchers used the Supercomputing Institute's resources to analyze three measures of blood pressure (systolic, diastolic, and pulse) in the San Antonio Family Heart Study. Given this large dataset and the power of the supercomputer, it was possible to consider models whose complexity had heretofore been thought impractical. Specifically, this approach computed an optimal model at each of the 400 markers, thus reducing the gene-gene complexity. This optimal model included gene-environment interactions, thus reducing environmental complexity.

  A secondary concern of this group was the comparison of methods in genetic epidemiology. The major problem in Genetic Epidemiology was the detection and location of genes that cause common diseases in humans, for example, heart disease, diabetes, obesity, cancer. It is commonly accepted that multiple genes cause these diseases and that these genes interact with each other and with the environment. Several competing methods attempt to solve this problem. The relative performance of these methods is largely unknown; close form solutions to problems do not exist, so that power analyses must be performed by simulation. However, simulation is difficult due to the problem of testing a large number of genetic models that might adequately represent the actual range of genetic architecture underlying these complex traits. Also, these methods are all family-based, but the optimal family type (sibling-parent, nuclear, extended, etc.) is largely unknown.

  This research used the supercomputer to determine the conditions under which each methodological approach is preferable. A broad range of genetic models and family structures was examined. For each model-family combination, the two most popular methodological approaches were used to detect and localize genes affecting a simulated trait. These two approaches fall largely into two areas: model-based analysis and model-free analysis (or parametric and non-parametric). The question-which method, under which conditions?-while important to this field, is so computationally daunting that it has previously gone largely unexplored.

  The group's research covered several aspects of the general research area. One focus, producing a genome-wide linkage analysis of blood pressure in Mexican-Americans, considered the genetic mechanisms that control variation in blood pressure level are largely unknown. One of the first steps in understanding these mechanisms is the localization of the genes that have a significant effect on blood pressure. The researchers performed genome scans of systolic (SBP) and diastolic blood pressure (DBP) on a population-based sample of families in the San Antonio Family Heart Study. A likelihood-based Mendelian model incorporating genotype-specific effects of sex, age, age², body mass index (BMI), and blood pressure (SBP or DBP, as appropriate) as covariates was used to perform two-point lodscore (Z) linkage on 399 polymorphic markers. Results showed that the genotype-specific covariate effects were highly significant for both SBP and DBP. Linkage results showed that a quantitative trait locus (QTL) influencing DBP was significantly linked to D2S1790 (Z=3.92, θ=0.00) and showed suggestive linkage to D8S373 (Z=1.92, θ=0.00). A QTL influencing SBP showed suggestive linkage to D21S1440 (Z=2.82, θ=0.00) and D18S844 (Z=2.09, θ=0.00). Without the genotype-specific effects in the model, the linkage to D2S1790 as not even suggestive (Z=1.33, θ=0.09), thus genotype-specific modeling was crucial in detecting this linkage. A comparison with linkage studies based in other population showed that the significant linkage to D2S1790 has been replicated at the same marker in the Quebec Family Study. The replicated significant linkage at D2S1790 may begin to establish the locations of the genes that significantly affect blood pressure across several human ethnic groups.

  Similar work was carried out on the linkage of blood pressure to D21S1440 in Mexican Americans. This project aimed to perform a genome scan to detect chromosomal regions that are linked to both SBP and DBP blood pressure in Mexican Americans. A random subset of 10 of families from the San Antonio Family Heart Study was genotyped for 399 polymorphic markers (approximately 441 individuals with marker data). A bivariate likelihood-based Mendelian model in which the two traits were SBP and DBP and incorporating sex-specific and genotype-specific effects of age, age², and BMI as covariates was evaluated on the full dataset. The Mendelian model assumed a single-locus with two alleles in Hardy-Weinberg equilibrium. Effects of high blood pressure medication were incorporated into the model by using a threshold value. The resulting model was used to perform two-point lodscore analysis on all markers.

  The results of this analysis showed that sex-specific and genotype-specific effects of age, age² were each significant (p<.005) for both SBP and DBP. For SBP, BMI had six-specific effects (p=.008) but not genotype-specific effects (p=.054). For DBP, BMI had neither sex-specific (p=.116) nor genotype-specific effects (p=.194), but did have a constant effect (p<.0001). The genome scan showed significant evidence for linkage of blood pressure to D21S1440 (lodscore=3.35, recombination fraction (θ) was 0.02). There was suggestive evidence for linkage to D8S1100 (lodscore=2.07, θ=0.13).

  As this was the first report of a significant linkage for both systolic and diastolic blood pressure in a population-based sample, the group expected to continue its work within this important research area.

  The researchers continued to examine blood pressure and expand the research scope to include other cardiovascular phenotypes, such as cholesterol. This expansion also included the NHLBI Family Heart Study.