Powerful extreme phenotype sampling designs and score tests for genetic association studies

Abstract

We consider cross-sectional genetic association studies (common and rare variants) where non-genetic information is available or feasible to obtain for N individuals, but where it is infeasible to genotype all N individuals. We consider continuously measurable Gaussian traits (phenotypes). Genotyping n < N extreme phenotype individuals can yield better power to detect phenotype-genotype associations, as compared to randomly selecting n individuals. We define a person as having an extreme phenotype if the observed phenotype is above a specified threshold or below a specified threshold. We consider a model where these thresholds can be tailored to each individual. The classical extreme sampling design is to set equal thresholds for all individuals. We introduce a design (z-extreme sampling) where personalized thresholds are defined based on the residuals of a regression model including only non-genetic (fully available) information. We derive score tests for the situation where only n extremes are analyzed (complete case analysis) and for the situation where the non-genetic information on N - n non-extremes is included in the analysis (all case analysis). For the classical design, all case analysis is generally more powerful than complete case analysis. For the z-extreme sample, we show that all case and complete case tests are equally powerful. Simulations and data analysis also show that z-extreme sampling is at least as powerful as the classical extreme sampling design and the classical design is shown to be at times less powerful than random sampling. The method of dichotomizing extreme phenotypes is also discussed.

Multi-omics Data Integration for Identifying Osteoporosis Biomarkers and Their Biological Interaction and Causal Mechanisms

“…First, our sample size is relatively small. However, our study is the largest so far in the bone field for multi-omics analyses and we applied an extreme phenotype sampling strategy with stringent inclusion and exclusion criteria, which is known to provide enhanced statistical power for association analysis compared with studies using comparable numbers of randomly sampled subjects (Bjornland et al, 2018). Moreover, data from different omics levels can provide complementary and inherent validation information with each other, and thus, integrating multi-omics data can partially compensate for the relatively small sample sizes (Hasin et al, 2017).…”

Impact of rare and common genetic variation in the Interleukin-1 pathway on human cytokine responses

“…Another method to increase power in rare variant association studies, is to use extreme phenotypes, for the purpose of enriching rare variants with strong(er) effects in these outlier individuals [37,38]. In this study, we used 1% extreme cytokine producers in a binary association and overlapped the results with our continuous tests, for the purpose of 1) characterization of stimulispecific mechanisms, and 2) providing an indication of direction.…”

Loci associated with variation in gene expression and growth in juvenile salmon are influenced by the presence of a growth hormone transgene

“…We employed an extreme phenotype sampling strategy to select fish for further genetic analysis; previous studies suggest that sampling from the tails of a phenotypic distribution has similar or increased power to detect genotype/ trait associations as sampling from the full population which allows us to avoid performing costly genomic analysis on a prohibitively large number of individuals [63][64][65]. The roughly 20% (numbers were adjusted slightly to accommodate a 96-well format) largest and smallest fish by weight from each experimental group were selected in descending and ascending rank order, respectively, for subsequent analyses: NT Large (n = 47; weight range 9.8-15.2 g), NT Small (n = 46; 4.5-7.0 g), T Large (n = 46; 15.0-22.0 g), and T Small (n = 47; 3.0-9.3 g).…”